The controversy began when Saralees Nadarajah and A. Ba published a commentary in Frontiers in Communications and Networks in 2025, claiming that a well-known analytical model for jitter, developed by Hamza Dahmouni, André Girard, and Brunilde Sansò in 2012, contained errors in four of its central formulas. The original 2012 paper, published in the Annals of Telecommunications, had derived expressions for the average jitter experienced by a tagged stream of traffic—say, all the packets belonging to one user’s video stream—as a function of the load on a network queue. Nadarajah and Ba proposed an alternative expression that they presented as more general, and they raised three specific claims against the original work.

In a response published on 14 July 2026 in the same journal, Dahmouni, of the Institut National des Postes et Télécommunications in Rabat, together with Girard and Sansò, both of Montréal, systematically dismantled each of those claims. Their rebuttal is a masterclass in how queuing theory conventions work, and it carries a lesson for anyone who has ever tried to generalize a formula beyond the assumptions under which it was derived. The stakes are not merely academic: jitter models feed directly into the design of voice-over-IP systems, video conferencing platforms, and real-time control networks, where a few milliseconds of unmodeled delay variation can mean the difference between intelligible speech and unintelligible noise.

The first claim made by Nadarajah and Ba was that the 2012 paper suffered from confusion about which queue was actually being investigated. The rebuttal authors point out that this claim is straightforwardly inconsistent with the text of the original paper. Dahmouni and colleagues state in the abstract, in Sections 1.1 and 1.3, and again at the beginning of Section 3 that they assume Poisson traffic and exponential service times. In queuing theory, those two assumptions define precisely what is known as an M/M/1 queue in Kendall’s notation, the standard shorthand introduced by David Kendall in 1951 and codified in Leonard Kleinrock’s foundational 1975 textbook. The first M stands for the memoryless, or Markovian, arrival process; the second M for the Markovian service process; and the 1 for a single server. There is no ambiguity, the rebuttal authors argue, about what system the formulas describe.

The second and more substantive claim concerned the role of a parameter that appears in the jitter expression. Nadarajah and Ba argued that the 2012 formula is valid only under a restricted condition, and they proposed an alternative expression, which they labeled Equation 2, as a general formulation. Their proposed formula was derived from Equation 7 of the original paper under three assumptions: that the arrival process is exponential with a given parameter, that the service time is exponential with another parameter, and that the transit time—the total time a packet spends in the system—is exponential with a third parameter. Crucially, Nadarajah and Ba treated this transit-time parameter as a quantity independent of the arrival and service parameters, and they claimed that setting it equal to a particular value recovers the original 2012 result as a special case of their more general expression.

It is exactly this independence assumption that the rebuttal authors identify as the fatal flaw. In an M/M/1 queue, the transit time is well known to follow an exponential distribution whose parameter is fully determined by the arrival rate and the service rate. The mean transit time equals one over the difference between the service rate and the arrival rate, a result that appears in Kleinrock’s 1975 textbook and in W. J. Stewart’s 2009 volume on probability and queuing. In other words, the transit-time parameter is not a free variable that can be chosen at will; it is a function of the traffic parameters, and the only admissible value is the one fixed by the arrival and service rates. Substituting that admissible value into Nadarajah and Ba’s Equation 2 yields, algebraically, the original 2012 formula. The proposed generalization, the rebuttal authors conclude, collapses back into the very expression it was supposed to correct.

The third claim was perhaps the boldest: that the proposed expression could be applied to queues other than M/M/1, for instance when either the arrival or the service process is not exponential. To test this, the rebuttal authors turned to their own 2016 work on non-Poisson traffic, published by Dbira, Girard, and Sansò in the Annals of Telecommunications, and examined the case of a G/M/1 queue—a system with a general, possibly non-exponential arrival process but exponential service times. In such a queue, the transit time is still exponential, but its parameter depends on the probability that an arriving packet must wait, which itself is defined as the unique root in the interval between zero and one of an equation involving the Laplace transform of the inter-arrival distribution. That waiting probability is not an independent parameter; it depends entirely on the parameters of the inter-arrival and service-time distributions. The resulting jitter expression, the rebuttal authors show, differs fundamentally from the formula proposed by Nadarajah and Ba.

To make the point concrete, the rebuttal authors considered an arrival process following a Gamma distribution, characterized by a shape parameter and a scale parameter, combined with exponential service times. In that setting, the queue is described by three independent parameters—one for the service process and two for the arrival process—and the transit-time parameter becomes a function of all three. Nadarajah and Ba’s Equation 2, by contrast, presents the jitter as a function of three supposedly independent parameters, but once the transit-time parameter is recognized as dependent, the expression effectively contains only two independent variables while the queue it purports to describe is characterized by three. It is not clear at all, the rebuttal authors write, how such an expression could be used to compute the jitter of a G/M/1 queue. The mismatch holds for any queue in which the total number of parameters describing the arrival and service processes exceeds two.

The rebuttal authors summarize their position in five points. They state clearly that their 2012 analysis was based on exponential inter-arrival and service times, that the resulting jitter formula follows directly and is consistent with the standard definition of jitter under those assumptions, and that the interpretation of the proposed expression as a generalization appears to rely on treating the transit-time parameter as independent when, in the M/M/1 setting used to derive it, that parameter is a function of the arrival and service rates. They further note that the derivation of the proposed expression relies explicitly on the exponential form of the arrival and service processes, so further justification would be required before applying it to non-exponential processes, and that the G/M/1 example shows the proposed expression differs from known results, suggesting its use beyond the M/M/1 case may be limited.

For the broader community, the exchange is a reminder that analytical elegance and analytical validity are not the same thing. A formula that looks more general because it contains more symbols is not necessarily more general in substance; if one of those symbols is secretly determined by the others, the extra freedom is illusory. Jitter remains one of the hardest performance metrics to model in closed form, which is precisely why the Dahmouni group has spent years building a family of models, from simple M/M/1 systems with Poisson traffic to G/M/1 queues that accommodate non-Poisson arrivals. The 2026 response, edited by Koffiqa Khan of The University of the West Indies and reviewed by Santosh I. Gore, closes the loop on the specific claims but leaves the field with a durable principle: generalizations must respect the dependencies that queuing theory imposes, or they risk rediscovering, in more complicated clothing, the formulas they set out to replace.

Subject of Research: A 2026 rebuttal in Frontiers in Communications and Networks defends a classic 2012 analytical model for network jitter, showing that a proposed generalization collapses back to the original M/M/1 formula once queuing theory dependencies are respected.

Article Title: Corrections to the article ‘On the analytical model for jitter’ by Nadarajah and Ba

Article References: Dahmouni, H., Girard, A., & Sansò, B. (2026). Corrections to the article ‘On the analytical model for jitter’ by Nadarajah and Ba. Frontiers in Communications and Networks, 7, Article 1803255. https://doi.org/10.3389/frcmn.2026.1803255

Image Credits: AI Generated

DOI: 10.3389/frcmn.2026.1803255

Keywords: jitter, queuing theory, M/M/1 queue, packet delay variation, communication networks, quality of service, VoIP, analytical modeling, Corrections, article, analytical, model

The research, led by Parag Sohoni of Rabindranath Tagore University in Bhopal together with colleagues at several Indian institutions, describes a routing protocol called Fuzzy-AODV. It grafts a lightweight fuzzy inference system onto the Ad-hoc On-demand Distance Vector protocol, a workhorse of mobile networking, and adds an adaptive route maintenance mechanism. The goal is to improve quality of service, meaning better packet delivery, higher throughput, lower delay, and less wasted control traffic, in networks whose topology can change from one moment to the next.

The problem the team set out to solve is fundamental. Classic routing protocols such as AODV and OLSR were designed for mobile ad hoc networks where nodes move far more slowly than drones do. In a FANET, a route that was reliable a second ago may already be broken, because the aircraft that formed it has flown out of radio range. Proactive protocols like OLSR, which periodically broadcast routing updates, generate enormous overhead when the network churns this fast. Reactive protocols like AODV, which only discover routes when data actually needs to be sent, waste less control bandwidth, but they still choose paths blindly, with no sense of whether a candidate route is about to fail or is already congested.

Fuzzy-AODV addresses this by giving the routing decision engine three inputs that describe the health of each candidate path: node mobility, channel utilization, and queue utilization. Mobility captures how fast and unpredictably a drone is moving, which determines how long a link through that node is likely to survive. Channel utilization measures how heavily the shared radio bandwidth is being used, flagging nodes whose neighborhood is approaching congestion. Queue utilization reflects how full a node’s memory buffers are, since a drone with a jammed queue will drop packets no matter how good its wireless link looks. The authors note that few existing fuzzy routing schemes combine all three of these parameters in a single decision engine, and fewer still pair them with real-time route maintenance.

The fuzzy inference system itself is a Mamdani-type design, the classic architecture of fuzzy control. Each input is described by triangular membership functions at three linguistic levels, low, medium, and high, and the output classifies a route as excellent, good, regular, or terrible. Because three inputs each with three levels yield twenty-seven possible combinations, the entire rule base consists of just twenty-seven predefined rules. A route with low mobility, low channel use, and low queue use is rated excellent; a node rated high on all three is terrible and is never used. Fuzzification converts raw measurements into linguistic variables, the inference engine evaluates the rules, aggregation merges their outputs, and defuzzification produces a crisp score that the protocol uses to pick the best path.

Crucially, this design requires no training data, no labeled datasets, and no iterative optimization. That stands in sharp contrast to newer artificial intelligence approaches to drone networking, such as deep reinforcement learning, graph neural networks, and federated learning, which can achieve strong performance but demand substantial computing power, frequent model updates, and heavy communication. Drones carry small processors and limited batteries, so a routing scheme that adds a constant, tiny computational overhead per node is far more practical. The authors show formally that evaluating a candidate forwarding node takes constant time, meaning the overall complexity of route discovery remains dominated by the underlying broadcast mechanism, roughly proportional to the number of drones in the network.

The adaptive route maintenance mechanism is the protocol’s second pillar. Rather than waiting for a route to fail completely before launching an expensive rediscovery process, Fuzzy-AODV continuously monitors mobility, channel use, and queue use along active paths. When the fuzzy system detects that route quality is degrading, it triggers local repair or switches to an alternate path before packets start dropping. The protocol also balances load across multiple routes, distributing traffic to prevent bottlenecks. The authors argue that this self-healing behavior indirectly improves every quality of service metric: fewer route rediscoveries mean less routing overhead, while better route integrity means higher packet delivery ratios and shorter delays.

To test the design, the team ran simulations in the NS-2 network simulator over a 1200 by 1200 meter area with a radio range of 550 meters, varying the number of drones from 50 to 200 and using random mobility patterns. Each scenario was run ten times with different random seeds, and the reported values are averages across trials. The authors are candid that they did not perform formal statistical significance testing, such as confidence intervals or hypothesis tests, and plan to add that analysis in future work.

The results were consistent across all densities. Fuzzy-AODV delivered at least ten percent more packets successfully than the comparison methods, OLSR and an adaptive FANET routing scheme, and the advantage held even at the highest node counts. The protocol also produced roughly five percent lower network delay, reflecting fewer retransmissions and route interruptions, and it generated less routing overhead by eliminating unsuitable forwarding nodes during route establishment. Throughput improved as well, and the gains grew with node density, suggesting that the fuzzy decision mechanism scales gracefully as more drones crowd the airspace. In a detailed analysis of seventeen sender-receiver connection pairs, nearly all links were rated regular, good, or excellent, with only two pairs falling into the terrible category.

The authors are careful about the limits of their work. The simulations assumed cooperative drones, reliable access to mobility and queue data, and no hardware failures, malicious nodes, or large-scale outages. Real deployments will face GPS localization errors, radio interference, signal fading around obstacles, adverse weather, and heterogeneous drone platforms, any of which could degrade routing decisions. The threshold values in the fuzzy rules were tuned to typical FANET conditions, so applications with very different mobility patterns, such as disaster relief or military surveillance, may need to adjust the membership functions. The team also notes that no explicit energy model was included, although reducing unnecessary control traffic and retransmissions should indirectly cut energy consumption. Future work includes testing on real drone hardware, extending the fuzzy module with residual energy and link stability, and exploring hybrid designs that combine lightweight fuzzy inference with adaptive learning. For now, the study makes a compelling case that sometimes the smartest artificial intelligence is the kind that fits in a drone’s tiny brain.

Subject of Research: A fuzzy inference-based routing protocol for flying ad hoc networks of unmanned aerial vehicles

Article Title: Design a novel routing protocol for flying ad hoc networks based on fuzzy inference to enhance quality of service

Article References: Sohoni, P., Shrivastava, S. S., Mannepalli, P. K., Chugh, S., & Rai, M. (2026). Design a novel routing protocol for flying ad hoc networks based on fuzzy inference to enhance quality of service. Discover Artificial Intelligence, 6(1), Article 1385. https://doi.org/10.1007/s44163-026-02165-4

Image Credits: AI Generated

DOI: 10.1007/s44163-026-02165-4

Keywords: FANET, UAV, fuzzy logic, AODV, routing protocol, quality of service, Mamdani inference, drone networks, packet delivery ratio, throughput, network delay, NS-2 simulation

The problem HAMRL tackles is deceptively simple to state but brutally hard in practice. An IoT environment is a shifting landscape of heterogeneous devices: sensors, actuators, gateways, and cloud services, each with different capabilities, response times, and reliability levels. A user request rarely maps onto a single service. Instead, it demands a composition, a carefully ordered chain of services that together fulfill the request while respecting Quality of Service constraints such as latency, availability, and cost. In a static network, engineers could hand-craft these chains. But IoT networks are anything but static. Devices join and leave, wireless links degrade, workloads spike unpredictably, and the composition that worked perfectly yesterday may collapse today. Scalability compounds the difficulty: as the number of devices grows into the hundreds or thousands, the space of possible service combinations explodes combinatorially, overwhelming classical optimization techniques.

Reinforcement learning has emerged as a natural candidate for this challenge. In reinforcement learning, an agent learns by trial and error, taking actions in an environment and receiving rewards or penalties that gradually shape its behavior toward better outcomes. The approach has produced landmark results, from game-playing systems that reached human-level control to adaptive traffic signal networks. Applied to service composition, a learning agent can, in principle, discover which service chains deliver the best quality without needing an explicit model of every device. But single-agent learning hits a wall in large IoT systems, because no single agent can observe or manage the entire network. The field therefore turned to multi-agent reinforcement learning, where each IoT node hosts its own agent and the agents must coordinate. Coordination, however, introduces its own pathology: when every agent is learning simultaneously, each one perceives the others as part of a constantly changing environment, and the learning process can become unstable, a phenomenon researchers describe as the moving target problem.

The Kashan team, Sara Rezaei, Salman Goli-Bidgoli, and Fereshte Dehghani, designed HAMRL around a well-known architectural compromise called centralized training with decentralized execution. The framework combines three ingredients: decentralized actors, a centralized critic, and a hierarchical attention mechanism. Each actor is a small neural network that lives with an individual IoT agent and decides which service to select based on what that agent can locally observe. During training, however, a centralized critic, a second neural network with access to global information, evaluates how good the joint decisions of all agents actually were, providing each actor with a far more stable learning signal than any agent could compute alone. This division of labor, rooted in the actor-critic family of algorithms first formalized in 2000, means the system can learn with global oversight but operate without it, which is essential for real deployments where a global view may not exist at runtime.

The genuinely novel component is the hierarchical attention mechanism. Attention, the computational idea that now underpins modern language models, allows a neural network to weigh the relevance of different pieces of information rather than treating them all equally. HAMRL applies this idea at two levels. At the lower level, each agent fuses its local context, the state of its own device and immediate neighborhood, with global contextual information about the wider network, learning which of the many available signals actually matter for the decision at hand. At the higher level, the framework organizes these attention weights hierarchically, so that agents can reason about relationships across different scales of the network, from a single sensor cluster up to the system as a whole. The practical effect is that an agent choosing a service does not drown in irrelevant data from thousands of distant devices; it learns to focus on the handful of peers and conditions that genuinely influence quality of service.

This fusion of local and global context is what the authors identify as the key to HAMRL’s improved scalability and adaptability. Because each agent’s decision module only attends to the most relevant information, the computational burden does not grow uncontrollably as the network expands. And because the attention weights are learned rather than fixed, the system can re-prioritize its focus when conditions change, for example when a popular service becomes overloaded or a new device with better latency joins the network. Adaptability of this kind is precisely what static composition methods and even earlier learning-based approaches have struggled to deliver under the dynamic conditions that define real IoT deployments.

To find out whether these design choices actually pay off, the researchers ran extensive experiments comparing HAMRL against four state-of-the-art multi-agent reinforcement learning baselines: MAAC, an actor-attention-critic method; COMA, a counterfactual multi-agent policy gradient algorithm; MAA2C, a multi-agent extension of the widely used advantage actor-critic method; and MAPPO, a multi-agent version of proximal policy optimization that has surprised researchers with its effectiveness in cooperative settings. The evaluation focused on three criteria: the quality of service achieved by the composed solutions, the stability of the learning process, and adaptability when environmental conditions shifted. Across all three, HAMRL significantly outperformed the competition, according to the study. The hierarchical attention mechanism appears to give the agents a clearer, less noisy picture of their environment, which translates into steadier learning curves and better final compositions.

The implications reach well beyond benchmark experiments. Service composition sits at the heart of smart infrastructure: intelligent buildings, industrial automation, healthcare monitoring, smart cities, and logistics networks all depend on chaining device-level services into coherent applications. A composition framework that adapts in real time, maintains quality of service under load, and scales to large device populations could make these systems more resilient and less dependent on human reconfiguration. The study also contributes to a broader conversation in machine learning about the role of attention in multi-agent systems, a topic that has attracted growing interest as researchers seek principled ways to tame the complexity of many-agent coordination. By structuring attention hierarchically rather than flatly, HAMRL offers a template that other domains, from drone flocking to traffic control, might borrow.

None of this means the problem is solved. The authors frame HAMRL as a promising solution for next-generation scalable and resilient IoT applications, a carefully chosen phrase that acknowledges the gap between controlled experiments and messy production networks. Real deployments raise questions the paper’s benchmarks cannot fully answer: how the framework behaves under adversarial conditions, how it handles security and privacy constraints in service selection, and how it integrates with existing IoT protocols and standards. The researchers also note that the data supporting their findings are available from the corresponding author upon request, inviting the community to scrutinize and extend the results. Still, the central lesson stands. In ecosystems where no single controller can exist, intelligence has to be distributed, and the hardest part is not making individual agents smart but making them coordinate. HAMRL’s answer, letting agents learn what to pay attention to, locally and globally, at multiple scales, is an elegant step toward IoT networks that quietly assemble themselves, service by service, while the rest of us just ask for coffee.

Subject of Research: Multi-agent reinforcement learning for quality-of-service-aware IoT service composition

Article Title: HAMRL: a hierarchical attention-based multi-agent reinforcement learning approach for IoT-based service composition

Article References: Rezaei, S., Goli-Bidgoli, S., & Dehghani, F. (2026). HAMRL: a hierarchical attention-based multi-agent reinforcement learning approach for IoT-based service composition. Cluster Computing, 29(15), Article 832. https://doi.org/10.1007/s10586-026-06596-7

Image Credits: AI Generated

DOI: 10.1007/s10586-026-06596-7

Keywords: Internet of Things, service composition, multi-agent reinforcement learning, hierarchical attention, quality of service, centralized critic, decentralized actors, actor-critic algorithms, scalability, adaptability, smart infrastructure, Cluster Computing

Federated learning, first popularized by researchers at Google in 2017, flips the conventional training pipeline on its head. Instead of moving data to the model, the model moves to the data. Each device—or client—trains a local copy of the algorithm on its own measurements and then transmits only the resulting parameter updates, typically small numerical gradients, to a coordinating server. The server averages these updates into a shared global model and sends it back down for another round of local training. Because raw sensor readings never leave the device, the approach reduces direct exposure of sensitive information while slashing the volume of traffic that must cross the network. For IoT deployments, where a single greenhouse or smart building may host dozens of constrained devices, that distinction is not a luxury but a prerequisite.

What makes the Moroccan team’s contribution distinctive is where the federation logic lives. The OneM2M standard is an internationally recognized service layer designed to sit between heterogeneous IoT devices and applications, providing common primitives for registration, data storage, subscription, and notification regardless of whether devices speak HTTP, MQTT, or CoAP. Rather than bolting federated learning onto an ad hoc middleware stack, FL-OneM2M embeds it into this operational backbone, turning OneM2M into an orchestration infrastructure for decentralized intelligence. The authors argue that this integration addresses a genuine gap: most federated learning research assumes idealized communication conditions, while most IoT standards work ignores machine learning entirely. By unifying the two, the framework promises both quality-of-service awareness and scalable learning in a single architecture.

The experimental setup drew on data generated from IoT agricultural deployments, collected through Azure IoT devices and processed via OneM2M-based data injectors. The raw sensor streams included humidity, temperature, environmental plant indicators, and a suite of agricultural growth metrics, partitioned across multiple simulated greenhouse clients. On top of this dataset, the researchers deployed logistic-regression clients and compared two canonical federated optimization algorithms. Federated Averaging, known as FedAvg, is the simplest and most widely used scheme: clients train locally and the server weights each update by client data volume. Federated Proximal, or FedProx, adds a proximal term to the local objective that penalizes clients whose local updates drift too far from the global model, a modification intended to stabilize learning when devices are heterogeneous or perform varying amounts of local work.

The results, reported for an independently and identically distributed—or IID—data setting, tell a nuanced story. FedAvg reached a final accuracy of 80 percent, a gain of 50 percentage points over its starting point, and achieved a higher peak accuracy than FedProx, which topped out at 75 percent. FedProx, however, exhibited smoother loss convergence, suggesting that its regularization term does damp oscillations even when it costs some final performance on this particular problem. The authors are careful to frame these numbers as trends observed under their specific conditions rather than universal verdicts. With a single model class, a single dataset, and at most ten clients, the study deliberately scopes its claims, and the paper discusses the resulting limitations—large-scale behavior, statistical heterogeneity across clients, and security—explicitly rather than glossing over them.

Perhaps the most striking findings concern communication efficiency, the resource that often determines whether federated learning is practical at the edge at all. Under the study’s test conditions, switching the transport protocol from HTTP to MQTT reduced the average round-trip time by 98.4 percent, and the system reached peak accuracy roughly 40 training rounds earlier. The explanation lies in protocol design. HTTP is a request-response protocol that carries substantial header overhead and requires a new exchange for every message, whereas MQTT is a lightweight publish-subscribe protocol built for constrained networks, using minimal headers and persistent connections. In a federated loop, where the server and clients exchange model updates hundreds of times, shaving milliseconds off every round compounds into a dramatic acceleration of the entire training process. For anyone deploying learning across battery-powered sensors, that difference can separate a viable system from an intractable one.

Scalability tests at small scale—three, five, and ten clients—added another practical insight: five clients produced the most stable convergence in the studied configuration. This counterintuitive result reflects a real tension in federated systems. With very few clients, each local update carries outsized weight, so a single noisy or poorly conditioned client can destabilize the global model. As the client count grows, averaging smooths out individual noise, but coordination overhead rises and, in real deployments, stragglers and dropouts multiply. The team also probed resilience by simulating clients dropping out mid-training and found that the system retained 80 percent accuracy under a 10 percent client-dropout rate, an encouraging signal that the OneM2M-based orchestration can absorb a degree of real-world unreliability without collapsing.

The choice of agriculture as the application domain is more than incidental. Smart farming is emerging as a natural proving ground for federated learning because farms are simultaneously data-rich and privacy-sensitive: yield forecasts, disease detections, and growth metrics carry commercial value, and farmers may be reluctant to surrender raw operational data to third parties. Recent work in the field has explored federated crop disease detection from images, model-pruned yield forecasting, and blockchain-assisted trust mechanisms, all pointing toward the same conclusion that collaborative intelligence is only acceptable to many stakeholders if the data stays local. FL-OneM2M fits squarely into this movement, demonstrating that a standardized service layer can serve as the connective tissue between field sensors and a learning pipeline without a central data lake.

The authors are refreshingly candid about what their study does not establish. Ten clients is a toy scale compared with the thousands or millions of devices a production IoT network might involve, and behavior at large scale—where communication bottlenecks, stragglers, and server load become dominant—remains untested. The IID assumption is similarly generous: real sensor networks are notoriously non-IID, with each greenhouse, factory floor, or city block exhibiting its own statistical fingerprint, a condition known to degrade federated averaging and motivate algorithms like FedProx, SCAFFOLD, and adaptive federated optimization in the broader literature. Security is another open front, since model updates themselves can leak information or be poisoned by malicious clients, and the paper positions these challenges as explicit future work rather than solved problems. The researchers also note that the underlying dataset was generated specifically for the study and contains sensitive operational information, so it is not publicly available, though processed or anonymized subsets can be requested from the corresponding author.

Even within those limits, the study offers a genuinely useful template. It shows that federated learning need not live in bespoke middleware but can ride on an open, interoperable standard that the IoT industry already deploys, and it quantifies how much protocol choice—MQTT over HTTP—matters to the speed of decentralized training. As edge hardware grows more capable and privacy regulation tightens worldwide, the pressure to move intelligence out of the cloud and into the network’s periphery will only intensify. Frameworks like FL-OneM2M suggest a path in which the same infrastructure that routes sensor data today becomes the stage on which devices learn together tomorrow, one carefully averaged model update at a time.

Subject of Research: Federated learning integrated with the OneM2M service layer for quality-of-service-aware, scalable machine learning in Internet of Things networks

Article Title: A novel federated learning approach for enhancing QoS and scalability in IoT networks

Article References: Et-Tousy, J., & Zyane, A. (2026). A novel federated learning approach for enhancing QoS and scalability in IoT networks. Cluster Computing, 29(14), Article 825. https://doi.org/10.1007/s10586-026-06633-5

Image Credits: AI Generated

DOI: 10.1007/s10586-026-06633-5

Keywords: federated learning, Internet of Things, OneM2M, edge intelligence, MQTT, quality of service, FedAvg, FedProx, smart agriculture, privacy-preserving machine learning, scalability, Cluster Computing

The core problem the researchers address is known as service composition, the task of assembling individual cloud services into a coherent workflow that fulfills a user request. A single application might depend on data storage, processing nodes, and specialized software components distributed across many servers, each combination carrying different costs, speeds, and reliability characteristics. Traditional approaches treat this as a static optimization puzzle: find the best arrangement, deploy it, and hope conditions hold. But conditions never hold. Demand fluctuates hour by hour, hardware fails unpredictably, and the optimal composition of one moment can become badly mismatched to reality the next. When that happens, operators face a painful trade-off between leaving a degraded system in place and paying the substantial cost of migrating services to new servers while customers are already waiting.

Shahab and his team, working across institutions in Canada, Iran, and Mexico, reframed this challenge as a sequential decision-making problem that a machine learning agent can learn to solve continuously. Their framework uses deep reinforcement learning, a family of techniques in which an artificial agent learns by acting, observing the consequences, and gradually improving its policy through repeated interaction with its environment. Rather than prescribing rules for when and how to move services, the agent discovers them. It learns to weigh the immediate pain of migration against the long-term benefit of a better configuration, developing a behavioral strategy that balances stability with adaptability, the two forces that pull cloud management in opposite directions.

What distinguishes this work from earlier reinforcement learning applications in cloud computing is the structure of its reward signal. The researchers designed a unified cost function that simultaneously captures three competing concerns: Quality of Service, the technical measure of how well the composed services perform for end users; migration costs, the operational expense and disruption caused by moving services between servers; and Service Level Agreement violations, the contractual penalties incurred when promised performance thresholds are breached. By folding all three into a single objective, the framework avoids the trap of optimizing one dimension while silently degrading another. An agent that migrates aggressively to chase marginal performance gains will feel the migration penalty; an agent that sits idle to avoid those penalties will accumulate SLA violations when conditions deteriorate. The learning process forces a genuine equilibrium.

The study does not bet on a single algorithm. Instead, the authors evaluated five advanced deep reinforcement learning methods within the same framework, subjecting each to the same dynamic environments of fluctuating demand and server failures. Among the contenders, Twin Delayed Deep Deterministic Policy Gradient, known as TD3, emerged as the strongest performer, achieving superior adaptability under dynamic conditions. TD3 belongs to a class of algorithms designed for continuous action spaces, where decisions involve fine-grained adjustments rather than discrete choices, and its characteristic twin-critic architecture helps suppress the overestimation of action values that can destabilize learning. Its victory in this setting suggests that the granular, continuous nature of cloud resource reallocation rewards algorithms built for precisely that kind of control problem.

To demonstrate practical relevance, the researchers grounded their framework in a case study drawn from one of the defining disruptions of recent memory: the pandemic-era surge in ventilator production. During that crisis, manufacturing networks depending on cloud services had to cope with explosive demand spikes, supply interruptions, and shifting priorities, all while maintaining the reliability that medical equipment production demands. The case study illustrates how a cloud network governed by the framework could dynamically recompose its services as conditions shifted, reallocating computational resources to keep production coordination running even as individual servers failed or demand patterns changed beyond anything the original configuration anticipated.

One of the study’s most actionable findings came from sensitivity analysis, a technique for probing how changes in individual parameters ripple through system performance. The analysis identified migration costs as the single most influential factor on overall performance. This result carries real weight for system designers. If moving services is cheap, an agent can afford to adapt frequently, chasing every fluctuation in demand with a responsive reconfiguration. If migration is expensive, frequent moves become self-defeating, and the optimal strategy shifts toward patience and tolerance of temporary suboptimality. The finding highlights that the balance between stability and adaptability is not an abstract philosophical question but a tunable engineering parameter, and that infrastructure decisions about migration infrastructure and virtualization overhead directly shape how resilient a cloud network can become.

The authors are candid about the scope of their contribution and its limits. The framework is algorithmically extensible, meaning its structure can accommodate different learning algorithms and be adapted to diverse cloud network settings, but they note that broader empirical scalability requires further testing on larger service and task networks. Real production clouds involve thousands of services, intricate dependency graphs, and adversarial conditions that laboratory simulations can only approximate. Validating that the learned policies scale gracefully, and that training remains tractable as the state space explodes with system size, remains an open frontier. The study also addresses a specific gap in the resilience literature: most prior work handles either demand fluctuations or resource failures, but rarely both simultaneously, which is precisely how real disruptions arrive.

The significance of this research extends beyond cloud computing into the broader question of how societies depend on invisible computational infrastructure. Healthcare systems, energy grids, logistics networks, and financial markets all ride on cloud services whose failure cascades into the physical world. The pandemic exposed how brittle tightly optimized systems can be when conditions swing violently, and a growing body of resilience research across supply chains, energy systems, and civil infrastructure has converged on the insight that adaptability must be designed in, not bolted on after the fact. A cloud network that continuously learns to reorganize itself under stress embodies that principle at the software layer, turning resilience from a static property of redundancy into a dynamic capability of learning.

For the engineers and researchers watching this space, the study offers a template worth studying closely: a unified cost function that resists single-metric myopia, a comparative evaluation across multiple state-of-the-art algorithms rather than a defense of one, and a sensitivity analysis that tells practitioners where their leverage lies. As deep reinforcement learning matures from game-playing demonstrations into operational infrastructure, work like this marks the transition, showing that the same techniques that mastered abstract games can be entrusted with keeping the digital backbone of critical services standing when the world refuses to cooperate. The cloud, it turns out, can learn to bend without breaking.

Subject of Research: Deep reinforcement learning for resilient dynamic cloud service composition under disruptions

Article Title: Dynamic cloud service compositions under disruptions: a deep reinforcement learning framework

Article References: Shahab, E., Rabiee, M., Eslami, A., Gholian-Jouybari, F., & Hajiaghaei-Keshteli, M. (2026). Dynamic cloud service compositions under disruptions: a deep reinforcement learning framework. Cluster Computing, 29(13), Article 765. https://doi.org/10.1007/s10586-026-06497-9

Image Credits: AI Generated

DOI: 10.1007/s10586-026-06497-9

Keywords: cloud computing, deep reinforcement learning, service composition, resilience, TD3, Quality of Service, Service Level Agreement, server failures, demand fluctuations, migration costs, cloud networks, machine learning

The stakes are enormous. Fifth-generation technology, commercialized in 2019, can connect roughly one million devices in an area of 0.38 square miles, deliver peak data throughput of 20 gigabits per second, and achieve minimum latency of one millisecond. Yet the survey’s authors argue that 5G shows restricted flexibility in dynamic contexts. Vehicular services demand ultra-low latency and high reliability simultaneously, and the varied, rigorous requirements of Vehicle-to-Everything communication, known as V2X, often exceed what 5G can guarantee. Beyond 5G and especially sixth-generation networks are anticipated to close that gap by tapping higher frequency spectrum, enabling more network capacity with substantially reduced latency.

The core innovation the survey examines is network slicing, a technique that divides a single physical network into multiple logically separate networks, each tailored to the needs of a different service. In a vehicular context, one slice might serve safety-critical messages between cars, another might carry infotainment streams for passengers, and a third might handle bulk sensor uploads for autonomous driving algorithms. Because each slice can be designed with its own quality of service guarantees, network slicing has significantly enhanced the growth of vehicular networks. The approach relies on softwarization technologies such as software-defined networking and network functions virtualization, which let operators carve, configure, and reconfigure slices on demand rather than through fixed hardware.

But slicing alone is not enough, and this is where the survey’s central argument emerges. Vehicles are, by definition, mobile, and high mobility is one of the defining characteristics of vehicular network members. When a car crosses from one cell to another, or from one network domain to a completely different one, its slice relationships must travel with it. Mobility management is the set of techniques that dynamically preserves these slice relationships and quality of service guarantees while users or devices move. In advanced vehicular networks, the authors emphasize, network slicing and mobility management are two key components that must collaborate to ensure quality of service. A slice that cannot follow the vehicle is a broken promise.

The difficulty is that vehicular surroundings are dynamic and varied in ways that defeat traditional, rule-based mobility management. Handover decisions, the moments when a connection transfers between base stations or between slices, depend on signal quality, traffic load, vehicle speed, and prediction of future position, all changing from second to second. Conventional schemes with fixed parameters, such as static handover thresholds and time-to-trigger values, struggle to adapt. The survey reviews a range of critical techniques for mobility management in 6G vehicular networks, including handover optimization, dual connectivity, positioning and localization methods, and predictive approaches that anticipate where a vehicle will be before the connection needs to move.

Enter deep reinforcement learning, a machine learning technique that combines deep neural networks with the trial-and-error logic of reinforcement learning. In this framework, an agent observes the state of its environment, takes an action, and receives a reward or penalty, gradually learning a policy that maximizes long-term benefit. Deep neural networks allow the agent to handle enormous, high-dimensional state spaces, such as the full radio and traffic conditions of a highway segment, that would overwhelm tabular methods. Algorithms surveyed in the paper include deep Q-networks and their variants, actor-critic families such as proximal policy optimization and soft actor-critic, and multi-agent extensions in which several learning agents cooperate to manage shared resources.

The applications documented in the survey span the entire vehicular networking stack. Deep reinforcement learning has been applied to radio access network slicing for cellular V2X, to resource allocation among competing slices, to task offloading in vehicular edge computing, and to handover decisions in heterogeneous networks. Multi-agent deep reinforcement learning approaches have been used for network slicing in vehicular communications, where distributed agents coordinate slice resources without a central controller. The technique has also been applied to beam management in millimeter-wave systems, trajectory planning for unmanned aerial vehicles that assist ground networks, and joint optimization of caching, computing, and radio resources. In each case, the promise is the same: adaptive, experience-driven decisions that outperform static rules in environments no engineer could fully anticipate.

Why does this matter beyond the laboratory? The advancement of transportation, the authors note, affects multiple facets of people’s lives, encompassing the economy, tourism, and healthcare. Reliable V2X communication underpins emergency vehicle prioritization, platooning, cooperative perception for autonomous driving, and remote monitoring applications. A network that drops a safety message during a handover is not merely inconvenient; it can be dangerous. By combining network slicing, which guarantees per-service resources, with intelligent mobility management, which guarantees continuity of those guarantees under motion, 6G vehicular networks aim to make connected driving dependable enough for safety-critical deployment. The survey also situates this within a broader 6G toolkit that includes intelligent reflecting surfaces, integrated sensing and communication, digital twin networks, and non-terrestrial components such as low Earth orbit satellites.

The authors are careful about limitations. They state that the use of deep reinforcement learning is a beneficial method to improve mobility management in sliced vehicular networks, but that further studies are required. Open challenges include the exploration problem in reinforcement learning, the training cost and sample efficiency of deep agents, the interpretability of learned policies, and the security of learning-based slicing systems, an area where federated learning approaches are being explored to detect attacks across sliced networks. Standardization efforts, including the ITU’s IMT-2030 framework for 6G and the Open Radio Access Network architecture, will shape how intelligent mobility management is actually deployed.

To encourage further work, the survey outlines potential future research objectives, ranging from end-to-end slice mobility across heterogeneous domains to explainable AI for resource management in vehicular network slicing. The authors hope the survey will help researchers understand in depth the concept of network slicing in vehicular networks and the impact of mobility management in sliced environments. For the rest of us, the message is simpler: the self-driving future depends not just on smarter cars, but on networks smart enough to learn, adapt, and never let go of a moving vehicle’s connection, no matter how fast the road unfurls beneath it.

Subject of Research: Deep reinforcement learning for mobility management in sliced 6G vehicular networks

Article Title: Applying deep reinforcement learning to manage mobility in a sliced 6G vehicular network: a survey

Article References: Amaira, A., & Zarai, F. (2026). Applying deep reinforcement learning to manage mobility in a sliced 6G vehicular network: a survey. Cluster Computing, 29(14), Article 820. https://doi.org/10.1007/s10586-026-06606-8

Image Credits: AI Generated

DOI: 10.1007/s10586-026-06606-8

Keywords: 6G, vehicular networks, network slicing, mobility management, deep reinforcement learning, V2X, 5G, handover, machine learning, quality of service, edge computing, beyond 5G

The algorithm, described in the journal Cluster Computing, is called Opposition-Based Firebug Tunicate Optimization, or OFTO. It is the brainchild of Chirag Chandrashekar, Pradeep Krishnadoss, Arun Kumar Sivaraman, Vijayakumar Kedalu Poornachary, Kong Fah Tee, Janakiraman Nithiyanantham and colleagues, and it addresses a problem that has dogged cloud schedulers for years: the delicate balance between exploring the search space and exploiting promising solutions once they are found. Metaheuristic algorithms — the family of nature-inspired techniques that includes particle swarm optimization, ant colony methods, and grey wolf optimizers — are widely used for scheduling because the underlying problem is combinatorial and grows explosively with the number of tasks and machines. But many of these algorithms converge prematurely, locking onto a mediocre schedule before they have surveyed the full landscape of possibilities.

OFTO tackles that weakness by fusing two existing swarm intelligence techniques, each with complementary strengths. The first is Firebug Swarm Optimization, an algorithm modeled on the mating behavior of firebugs, which excels at local search — refining a candidate solution in its immediate neighborhood. The second is Tunicate Swarm Optimization, inspired by the jet-propulsion movement and swarming of marine tunicates, which provides robust global search, allowing the population to leap across distant regions of the solution space. By combining the two, the researchers created a hybrid whose population can both probe widely and drill deeply, avoiding the twin failure modes of wandering forever without converging and of settling too early on a local optimum.

The third ingredient is what gives the algorithm its name’s prefix: opposition-based learning. The idea, first proposed in the mid-2000s, is deceptively simple. When searching for a good solution, it often pays to evaluate not just a candidate but also its mathematical opposite — a point reflected across the center of the search space. If the opposite point happens to be better, the algorithm jumps there; if not, the comparison still enriches the population’s diversity. In OFTO, every iteration evaluates both current and opposite candidate solutions, which the authors say ensures comprehensive coverage of the search space while maintaining exploitation efficiency, accelerating convergence and helping the algorithm escape local optima that would otherwise trap less adaptable schedulers.

Crucially, OFTO is not just a generic optimizer repurposed for the cloud. The researchers designed it as what they call a Deadline Sensitive Task Scheduler, tuned to the quality-of-service parameters that matter most in commercial cloud environments: execution cost, makespan — the total time to finish all tasks — and the speed at which a good schedule is found. In practice, that means the algorithm assigns each incoming task to the virtual machine best positioned to complete it within a constrained time frame while consuming minimal computational resources. For cloud providers, whose profit margins depend on keeping utilization high and service-level agreements intact, that combination of speed, cost, and deadline awareness is precisely the trade-off they need to manage.

To test their scheduler, the team ran extensive simulations against a lineup of state-of-the-art competitors, including a fault-tolerant trust-based Harris Hawks deep reinforcement learning algorithm known as FTTHDRLA, a modified transfer-function-based binary particle swarm optimizer called MTF-BPSO, the Rider Cuckoo Optimization Algorithm, and the two parent algorithms, Tunicate Swarm Optimization and Firebug Swarm Optimization, on their own. The results were striking. OFTO improved overall performance by 7.96 percent over FTTHDRLA, 14.98 percent over MTF-BPSO, 18.50 percent over RCOA, 24.75 percent over TSO, and 31.71 percent over FSO. The fact that it beat both of its own parent algorithms suggests the hybridization and the opposition mechanism each contribute something the individual components lack.

The authors did not rely on raw performance numbers alone. They subjected their results to a battery of statistical significance tests — the T-test, the Wilcoxon signed-rank test, and the Friedman test — which are standard tools for establishing that one optimizer’s advantage over another is not a fluke of a particular random seed or workload. The tests, they report, confirm the robustness and reliability of the proposed algorithm across multiple metrics, lending the headline improvements a firmer footing than a single favorable benchmark would provide.

The experimental setup itself reflects the practical constraints of cloud research. Because cloud service providers enforce strict data confidentiality policies, real user workload data is notoriously difficult for researchers to obtain, and the study was conducted without external funding, putting real-time cloud execution beyond the project’s scope. Instead, the team used an open-source benchmark, the GoCJ Google Cloud Jobs Dataset hosted on the Mendeley Data repository, which is derived from real-world Google cluster traces and MapReduce logs. That gives the simulations a credible anchor in genuine production workloads, even though the experiments were run in simulation rather than on a live commercial cloud — a caveat that readers should keep in mind when weighing the reported gains.

The work arrives amid a flourishing of hybrid metaheuristics for cloud scheduling. Recent years have seen krill herd algorithms for fog-cloud workflow scheduling, chameleon and remora search for handling scheduling uncertainty, improved wild horse optimization with Levy flight, fuzzy quantum atom search optimizers, beluga whale optimization with ring topologies, and adaptive chaotic reverse learning-enhanced reptile search, among many others. The proliferation reflects both the difficulty of the scheduling problem and the hunger of the cloud industry for any edge in efficiency. Against that crowded backdrop, OFTO’s distinguishing features are its explicit deadline sensitivity and its systematic use of opposition-based learning to keep population diversity high throughout the search.

For the researchers, the next steps will likely involve testing the algorithm under more dynamic conditions — fluctuating workloads, machine failures, and heterogeneous resource types — and, eventually, deployment on real infrastructure. For the rest of us, the study is a reminder that some of the most consequential computation in the modern world happens before our tasks even begin to run. The difference between a cloud data center that hums and one that thrashes comes down to algorithms like these: silent, insect-inspired matchmakers deciding, millions of times a second, where the world’s digital work should go. If OFTO’s reported gains hold up outside the simulator, the firebug and the tunicate may have just earned a small but real place in the plumbing of the internet.

Subject of Research: Hybrid opposition-based swarm intelligence for deadline-sensitive task scheduling in cloud computing

Article Title: Efficient task scheduling in cloud computing environment using opposition-based firebug tunicate optimization

Article References: Chandrashekar, C., Krishnadoss, P., Sivaraman, A. K., Poornachary, V. K., Tee, K. F., & Nithiyanantham, J. (2026). Efficient task scheduling in cloud computing environment using opposition-based firebug tunicate optimization. Cluster Computing, 29(14), Article 819. https://doi.org/10.1007/s10586-026-06561-4

Image Credits: AI Generated

DOI: 10.1007/s10586-026-06561-4

Keywords: cloud computing, task scheduling, swarm intelligence, metaheuristics, opposition-based learning, makespan, quality of service, virtual machines, Firebug Swarm Optimization, Tunicate Swarm Optimization, deadline scheduling, Cluster Computing

A research team led by P. Suma of Telangana Social Welfare Residential Degree College in Warangal, India, together with colleagues from National Kaohsiung University of Science and Technology in Taiwan, Sreenidhi Institute of Science and Technology in India, South East Technological University in Ireland, and Sri Padampat Singhania University in India, has introduced a hybrid nature-inspired optimization technique called the Greater Cane Mayfly Algorithm, or GCMA. The method is designed to select alternative paths and optimize network resources so that quality of service and communication security improve simultaneously. The work, published in September 2026, reports measurable gains across the standard performance metrics that network engineers use to judge routing protocols.

The inspiration behind the algorithm draws on two strands of the metaheuristic literature. Hybrid optimization methods frequently combine the exploration strength of one biological metaphor with the exploitation strength of another, and the GCMA follows this pattern by merging ideas associated with the greater cane rat algorithm, a recent swarm-based metaheuristic, with the mayfly optimization approach, which models the short-range mating behavior and long-range swarming flight of mayflies. In optimization terms, this means the algorithm balances broad searches across the solution space, which help it avoid getting trapped in local optima, with fine-grained refinement around promising candidates, which helps it converge quickly. Applied to routing, each candidate solution represents a potential path through the network, and the fitness of that candidate is judged against multiple criteria rather than a single measure.

Those criteria are at the heart of why the work matters. Traditional routing protocols in MANETs, such as the well-known ad hoc on-demand distance vector family, typically select paths based on hop count, meaning the route with the fewest intermediate nodes wins. That approach is simple and fast, but it ignores the practical realities of battery-powered wireless devices. A short path through nodes with depleted batteries, congested queues, or poor link quality will fail quickly, forcing the network to rediscover routes repeatedly and wasting scarce energy. The GCMA instead evaluates alternative paths by considering quality of service parameters, queue length at each intermediate node, available bandwidth, and remaining energy resources together. By weighing all of these factors simultaneously, the algorithm can steer traffic away from overloaded or fragile nodes and toward routes that are likely to remain viable.

The mechanics of the proposed routing framework unfold across several coordinated stages. During route discovery, source nodes flood route request messages through the network to find candidate paths toward a destination. When route reply packets travel back along those candidate paths, the scheme measures the probability that each route is stable, using information gathered by the intermediate nodes that the replies traverse. This stability estimation allows the source node to rank routes before committing traffic to them, and the overall procedure terminates once the source has gathered sufficient information. Route maintenance then monitors the active path, and when a link breaks because a node moves away or runs out of power, the alternative path selection mechanism kicks in. Rather than triggering a complete rediscovery cycle, the system can switch to a pre-ranked backup route, minimizing interruption to ongoing communication sessions.

Security considerations motivate much of this design. Because MANETs are decentralized and often deployed in dynamic or hostile environments, including disaster zones, battlefields, and temporary event venues, they are vulnerable to threats such as data tampering and eavesdropping. A routing layer that blindly trusts any available path can inadvertently deliver sensitive traffic through compromised or unreliable nodes. By incorporating stability probabilities and resource-aware selection into the routing decision, the framework reduces the chances of traffic being routed through nodes that exhibit abnormal behavior or deteriorating conditions. The authors position the technique as a step toward safe communication in these decentralized systems, combining quality of service enhancements with mechanisms that make manipulation of the routing process harder for an adversary.

The reported results quantify the gains. In the team’s evaluation, the GCMA achieved a normalized routing overhead of 20.365, a throughput of 9.321 megabits per second, a packet delivery rate of 83.680 percent, and an end-to-end delay of 0.254 seconds, with the delay figure representing the lowest value recorded in the study. Each of these numbers speaks to a different aspect of network health. Throughput measures how much useful data the network can deliver per unit of time, packet delivery rate captures the fraction of transmitted packets that actually reach their destination, end-to-end delay reflects how long packets take to traverse the network, and normalized routing overhead indicates how much control traffic is required per unit of delivered data. Low overhead matters especially in MANETs, where every control message competes with user data for the same wireless channel and drains the same batteries.

The study situates itself within a rich body of prior work on multipath and energy-aware routing in ad hoc networks. Earlier research has explored energy-centric swarm approaches such as tunicate-inspired algorithms, bandwidth-aware adaptive multipath schemes, cooperative medium access control designs to extend network lifetime, trust-based topology-hiding protocols, and bio-inspired routing techniques drawn from ant colonies and hybrid metaheuristics like the mayfly harmony search and the multi-objective M-LionWhale model. Cross-layer designs that share information between physical, medium access, and network layers have also been proposed to improve energy efficiency. What distinguishes the current contribution is the specific hybrid pairing of the greater cane rat and mayfly metaphors, together with an integrated pipeline that ties stability estimation, queue and bandwidth awareness, and energy considerations directly into both primary and alternative path selection rather than treating them as separate optimization problems.

The practical implications extend to any scenario where infrastructure is absent or destroyed. Emergency response teams operating after earthquakes, military units in the field, vehicular networks on highways, and internet-of-things deployments in remote regions all depend on networks that can reconfigure themselves on the fly. In these settings, a routing layer that can anticipate link failures and switch to backup paths without flooding the network with control messages translates directly into longer battery life, more reliable video and voice communication, and better resilience under stress. The resource-aware path scoring also aligns with broader sustainability goals in networking, since minimizing wasted transmissions reduces the energy footprint of the entire system.

Limitations and open questions remain, as with any simulation-based study. The authors note that no datasets were generated or analyzed beyond the study’s own evaluation environment, and real-world validation across heterogeneous hardware, varied mobility models, and adversarial conditions would be the natural next step for this line of research. Scaling the algorithm to very large networks and assessing its behavior under sophisticated attack models will also matter for adoption. Nevertheless, the work offers a concrete demonstration that hybrid bio-inspired optimization can push MANET performance forward on multiple fronts at once, and it adds a new entry to the growing catalog of nature-derived strategies that computer scientists are using to tame systems that never stop moving. For a field in which every dropped packet and drained battery carries a cost, an algorithm that borrows the collective intelligence of swarming insects may prove a fitting guide.

Subject of Research: Nature-inspired hybrid optimization for routing and resource management in mobile ad hoc networks

Article Title: A Novel Hybrid Greater Cane Mayfly Optimization Algorithm for Alternative Path Selection and Resource Optimization Strategies in MANET

Article References: Suma, P., Shieh, C. S., Hussain, M. A., Tabassum, M., & Chakrabarti, P. (2026). A Novel Hybrid Greater Cane Mayfly Optimization Algorithm for Alternative Path Selection and Resource Optimization Strategies in MANET. Mobile Networks and Applications. https://doi.org/10.1007/s11036-026-02524-6

Image Credits: AI Generated

DOI: 10.1007/s11036-026-02524-6

Keywords: MANET, routing protocols, swarm intelligence, metaheuristics, quality of service, multipath routing, energy efficiency, network security, path selection, throughput, wireless networks, optimization

The research, carried out by Sahani Pooja Jaiprakash, Tapas Badal and Naween Kumar at Bennett University in Greater Noida, India, addresses a problem that has stubbornly resisted classical optimization. Cloud workflows are not simply piles of independent jobs; they are directed structures in which tasks depend on the outputs of other tasks, priorities shift, and the machines available to execute them differ wildly in processing power, energy profile and price. Traditional heuristics such as list-scheduling algorithms can produce good solutions for well-behaved cases, but the authors argue they tend to be inflexible, struggling to adapt to time-varying workloads and often overlooking task priorities and inter-task dependencies altogether.

The team’s answer is a framework they call RL-LBD, short for Reinforcement Learning with Long Short-Term Memory and DBSCAN. The name hints at its three cooperating components. The first is a two-stage Long Short-Term Memory network, a type of recurrent neural architecture designed to retain information across sequences. In the RL-LBD pipeline, the LSTM serves as the scheduler’s memory: it encodes the dependency structure of incoming tasks and the profiles of the machines that might execute them, so that the system does not treat each scheduling decision as if the world had just been created. That matters because in a real workflow, scheduling task forty without knowing that it blocks tasks forty-one through forty-five is a recipe for idle processors and missed deadlines.

The second component tackles scale. A scheduler facing thousands of tasks cannot sensibly reason about each one individually, so the framework uses DBSCAN, a density-based clustering algorithm, to group tasks with similar characteristics before any ordering decision is made. DBSCAN is a natural fit here because, unlike k-means-style methods, it does not require the number of clusters to be specified in advance and can label outliers, both useful properties when workload shapes are unknown and irregular. By collapsing the decision space into clusters of related tasks, the scheduling problem becomes tractable without discarding the structure that makes workflows hard in the first place.

On top of this memory-and-clustering substrate sits the third component: a Deep Q-Network agent, the same family of algorithm that famously learned to play Atari games from raw pixels. The DQN learns a policy that dynamically prioritizes the clustered tasks, effectively deciding which group of work should move to the front of the queue under current conditions. A greedy task allocation strategy then assigns the prioritized tasks to specific virtual machines, choosing placements that minimize three objectives simultaneously: makespan, the total time to finish the entire workflow; energy consumption; and execution cost. The greedy allocation is deliberately simple, because the heavy lifting of judgment has already been done by the learned prioritization policy.

Reinforcement learning is well suited to this setting for a structural reason. Unlike supervised learning, which needs labeled examples of correct answers, an RL agent learns by acting: it schedules tasks, observes the consequences in its reward signal, and adjusts. Cloud scheduling is precisely a sequential decision problem, where an early choice constrains everything that follows, and where the environment, fluctuating workloads, changing machine availability, is non-stationary. Earlier efforts in this direction, from Q-learning-based schedulers to deep Q-network approaches for multi-objective workflow scheduling, demonstrated the promise but often treated tasks as independent entities or required retraining when workflows changed. The LSTM component in RL-LBD is the authors’ mechanism for injecting temporal and structural awareness into what would otherwise be a memoryless agent.

The reported results are where the framework earns its headline. In simulations, RL-LBD yielded improvements in energy efficiency, makespan and processing cost compared with the most recent competing approaches. Crucially, the authors did not stop at point estimates. They applied one-way analysis of variance and Fisher’s Least Significant Difference post hoc tests to confirm that the performance improvements were statistically significant across all evaluated scheduling metrics. In a field where claimed gains sometimes evaporate under statistical scrutiny, the formal significance testing lends the comparison more weight than a simple bar chart would.

The framework also showed versatility across different workflow types, which the authors interpret as evidence of its ability to enhance quality of service, improve data center efficiency and reduce operational costs. Notably, they report that the system demonstrated competitive performance in a real-world heterogeneous cloud environment, not merely in synthetic benchmarks, and that it can answer basic scheduling queries in a multi-layered, dynamic way. That last capability points toward a future in which schedulers are not static policies but interactive services that cloud operators can query about trade-offs between speed, cost and energy in real time.

The broader context makes the work timely. Data centers already consume a substantial share of global electricity, and that share is growing as AI workloads multiply. Scheduling is one of the few levers that improves all three of the cloud’s most pressing metrics, latency, energy and money, without new hardware. The literature the authors build on reflects a decade of escalating ambition: hybrid evolutionary algorithms such as HEPGA and deadline-budget-constrained ant colony optimization, multi-objective methods based on artificial bee colonies and genetic algorithms, and a steady march of reinforcement learning schedulers from early random-task schemes through parallel SARSA agents to meta-reinforcement learning approaches for multi-workflow environments. RL-LBD’s contribution is architectural: it shows that memory, clustering and value-based reinforcement learning can be composed so that each covers the others’ weaknesses.

There are, of course, the usual caveats that accompany simulation-driven scheduling research. The abstract does not report specific percentage improvements or the exact workflow benchmarks used, and the authors state that no datasets were generated or analysed during the study, meaning independent replication will depend on the community reconstructing comparable test conditions. The interplay between the LSTM’s memory horizon and the DBSCAN clustering granularity also raises questions about how the framework behaves under extreme scale or highly bursty arrivals, questions the full paper presumably explores. Still, the combination of statistical rigor, multi-objective optimization and demonstrated performance in a heterogeneous setting positions RL-LBD as a credible step toward schedulers that treat cloud workflows not as static puzzles but as living sequences to be learned. If the gains hold at production scale, the payoff could arrive quietly, in faster jobs, cooler server rooms and smaller invoices, exactly the kind of invisible victory a good scheduler is supposed to deliver.

Subject of Research: Adaptive scheduling of cloud computing workflows using LSTM-enhanced deep reinforcement learning to minimize makespan, energy consumption and execution cost

Article Title: Adaptive Cloud Workflow Scheduling via LSTM-Enhanced Deep Reinforcement Learning

Article References: Jaiprakash, S. P., Badal, T., & Kumar, N. (2026). Adaptive Cloud Workflow Scheduling via LSTM-Enhanced Deep Reinforcement Learning. Mobile Networks and Applications. https://doi.org/10.1007/s11036-026-02553-1

Image Credits: AI Generated

DOI: 10.1007/s11036-026-02553-1

Keywords: cloud computing, workflow scheduling, deep reinforcement learning, LSTM, DBSCAN clustering, Deep Q-Network, energy efficiency, makespan, virtual machines, quality of service, greedy task allocation, heterogeneous clouds

The stakes could hardly be higher. Every video stream, industrial sensor reading, autonomous vehicle decision, and healthcare monitoring alert now traverses a layered computing infrastructure in which computation may happen in a hyperscale data center, a regional fog node, or a device at the very edge of the network. Task scheduling, the process of deciding which computing resource will execute which job and when, directly determines latency, throughput, cost, and energy consumption. Load balancing, its close companion, spreads work evenly across available resources to prevent bottlenecks and failures. When either process falters, the consequences ripple outward: service level agreements are violated, users experience lag, servers idle wastefully or overheat, and in safety-critical applications such as connected vehicles or remote surgery, delays can be dangerous. The survey emphasizes that quality of service, resource efficiency, and overall system performance all hinge on how well these two intertwined problems are solved.

What distinguishes this survey from its predecessors is its insistence on a unified, comparative lens. Earlier reviews, the authors note, frequently examined cloud, fog, and edge environments in isolation, obscuring the trade-offs that emerge when workloads migrate between layers. By contrast, the new analysis traces how the optimal scheduling strategy shifts with position in the continuum. In the cloud, where resources are abundant but network latency to end users is significant, meta-heuristic algorithms dominate the literature. These nature-inspired optimizers, including genetic algorithms, particle swarm optimization, ant colony optimization, artificial bee colony methods, grey wolf optimizers, whale optimization algorithms, and cuckoo search, iteratively explore vast search spaces to find near-optimal assignments of tasks to virtual machines. The survey catalogues their underlying mechanisms in detail, from the mutation and crossover operators borrowed from biological evolution to the physics-inspired strategies modeled on phenomena such as simulated annealing and gravitational search.

In fog and edge environments, the picture changes dramatically. Resources are scarce, geographically dispersed, and volatile; network conditions fluctuate; and decisions must often be made in milliseconds. Here, the survey finds, the literature increasingly leans on artificial intelligence and machine learning, particularly reinforcement learning, in which an agent learns a scheduling policy through trial-and-error interaction with its environment, and deep reinforcement learning, which uses deep neural networks to handle enormous state spaces that would overwhelm tabular methods. The reviewed studies span Q-learning approaches for energy-efficient cloud scheduling, deep Q-network variants with improved exploration strategies, multi-agent deep reinforcement learning for distributed fog systems, soft actor-critic methods for task offloading in the internet of vehicles, and meta-reinforcement learning techniques designed to adapt rapidly to unseen edge scenarios. Applications range from smart factories and smart cities to healthcare management systems, where prioritizing critical patients over routine workloads can be a matter of life and death.

The technical anatomy of these learning-based schedulers is one of the survey’s richest contributions. In a typical deep reinforcement learning formulation, the state captures the current load of servers, queue lengths, task characteristics, and network conditions; the action space represents possible task-to-resource assignments or offloading decisions; and the reward function encodes objectives such as minimized makespan, reduced energy consumption, lower latency, or fewer service level agreement violations. Because real deployments rarely optimize a single metric, multi-objective formulations have proliferated, often combining reinforcement learning with meta-heuristics in hybrid schemes. One reviewed study, for example, coupled an artificial bee colony algorithm with reinforcement learning to balance competing objectives in cloud environments, while others integrated fuzzy logic to handle the uncertainty inherent in fog-cloud workflow scheduling. The authors also document the role of federated reinforcement learning, which allows edge nodes to learn collaboratively without centralizing sensitive data, and long short-term memory networks used for proactive scheduling of streaming applications before demand spikes materialize.

Measuring success in this field is itself a subtle science, and the survey devotes substantial effort to standardizing the metrics by which scheduling algorithms should be judged. Its appendix enumerates formal definitions for a comprehensive battery of indicators: makespan, the completion time of the last task in a batch; throughput, the number of tasks completed per unit time; response time and turnaround time; resource utilization; degree of imbalance; energy consumption; monetary cost; migration time; reliability, defined as the ratio of successful executions to total executions; availability; and service level agreement violation, computed from host active time and performance degradation. By assembling these definitions in one place, the authors provide researchers with a common vocabulary, addressing a long-standing problem in which studies claiming improvements often measure incomparable quantities across incomparable testbeds.

That testbed problem receives its own treatment. The survey reviews the simulation ecosystems on which the field depends, from the venerable CloudSim and its extensions such as CloudSim Plus and CloudAnalyst, to iFogSim and FogNetSim++ for fog scenarios, PureEdgeSim for the full cloud-fog-edge spectrum, and GreenCloud, iCanCloud, DCSim, and SCORE for energy-aware data center modeling. The authors classify these tools by environment, features, implementation language, open-source status, and graphical interface, noting that reproducibility and cross-paradigm evaluation remain weak points. Many studies validate their algorithms only within a single simulator tuned to a single layer, making it difficult to know whether a scheduler that excels in a cloud testbed will survive contact with the unpredictable wireless links and device heterogeneity of a real edge deployment.

From this synthesis, the survey distills a set of critical findings and uncomfortable gaps. Performance trade-offs emerge everywhere: meta-heuristics can approach optimal solutions but may converge slowly or demand heavy computation; reinforcement learning agents adapt dynamically but require extensive training and can behave unpredictably under distribution shift. More fundamentally, the authors identify the lack of cross-layer scheduling metrics as a key deficiency, since no standard measure exists to evaluate a decision that spans cloud, fog, and edge simultaneously. They also flag the insufficient handling of dynamic network conditions, pointing out that many algorithms assume stable connectivity even as real-world edge systems contend with fading links, mobility, and congestion. The absence of unified benchmarks spanning the full continuum, and the scarcity of studies addressing security and privacy constraints in scheduling decisions, round out a research agenda that the authors lay out for the community.

The future directions the survey champions read like a blueprint for the next decade of distributed computing. AI-driven adaptive scheduling, in which schedulers continuously learn and re-learn as workloads and network conditions evolve, stands at the forefront, alongside energy-aware optimization that treats power consumption as a first-class objective rather than an afterthought. The authors envision deeper integration of meta-learning so that scheduling agents can generalize across environments, wider adoption of federated approaches that preserve data sovereignty at the edge, and hybrid architectures that combine the global search power of meta-heuristics with the real-time responsiveness of learned policies. As billions of devices join the internet of things and applications from autonomous transport to industrial automation demand guaranteed latency, the survey argues, the systems that manage computation must themselves become intelligent. This work, published as volume 29, article 738 of Cluster Computing, positions itself as a foundational resource for the researchers and practitioners who will build that intelligent infrastructure, offering them a single map of a field that has, until now, been scattered across three separate territories.

Subject of Research: Task scheduling and load balancing techniques across cloud, fog, and edge computing systems

Article Title: Task scheduling and load balancing in cloud, fog, and edge computing systems: a survey

Article References: Zorgati, R., Hamdi, H., & Omri, M. N. (2026). Task scheduling and load balancing in cloud, fog, and edge computing systems: a survey. Cluster Computing, 29(13), Article 738. https://doi.org/10.1007/s10586-026-06513-y

Image Credits: AI Generated

DOI: 10.1007/s10586-026-06513-y

Keywords: task scheduling, load balancing, cloud computing, fog computing, edge computing, reinforcement learning, deep reinforcement learning, metaheuristics, quality of service, energy optimization, internet of things, artificial intelligence