Music-Inspired Algorithm Suggests Two Small Dams Could Rival One Giant Reservoir

A pair of modest reservoirs, orchestrated by an algorithm inspired by the way jazz musicians improvise, may be able to do the work of a giant dam. That is the provocative conclusion of a new study published in the journal Water Resources Management, in which Ahmed A. Al-Taey and Yousif H. Al-Aqeeli of the University of Mosul applied the harmony search algorithm to a proposed multi-reservoir system on the Greater Zab river along the border between Iraq and Turkey. Their results suggest that two smaller storage dams, operated under optimized policies, could produce annual hydropower approaching that of the far larger Bekhma reservoir, whose construction in Iraq is set to resume. The finding carries real weight for a country grappling with water scarcity, energy shortages and the geopolitics of transboundary rivers.

The study set out to determine operational policies for a multi-reservoir system composed of two single storage systems, referred to as Zab 1 and Zab 2, by maximizing annual hydropower generation. Rather than relying on rule curves drawn from historical practice or engineering intuition, the researchers turned to formal optimization, the mathematical craft of finding the best possible set of decisions under a given set of constraints. Reservoir operation is a classic optimization problem: releases must balance electricity production against storage targets, flood control and downstream demands, all while future inflows remain uncertain. Choosing when and how much water to release through turbines is therefore a decision problem of genuine complexity, one that grows harder still when multiple reservoirs in a cascade must be coordinated.

The tool the authors chose was the harmony search algorithm, a metaheuristic first proposed by Zong Woo Geem and colleagues in 2001. Its inspiration is deliberately unconventional. When a jazz musician seeks a pleasing harmony, they try out combinations of notes from memory and improvisation, gradually refining toward the best chord. The algorithm mimics this process: candidate solutions are stored in a harmony memory, new solutions are generated by drawing values from that memory with a certain probability, adjusting them slightly, or introducing entirely fresh random values, and promising candidates replace weaker ones over successive iterations. Because it handles continuous and discrete variables alike and imposes few restrictions on the objective function, harmony search has found a devoted following in water resources engineering, where objective functions are often nonlinear and the search landscape is riddled with local optima that trap simpler methods.

Harmony search is not new to reservoir problems. Previous work, including a well-known 2015 application to reservoir operation optimization by Bashiri-Atrabi and colleagues, demonstrated that the algorithm could find high-quality release schedules for single reservoirs, and other studies have extended it to multiple dam systems and compared it with rival metaheuristics. What distinguishes the new study is its combination of scale, setting and comparison. The researchers constructed three distinct harmony search models: the first and second represented the two reservoirs of the proposed multi-reservoir system individually, while the third formulated the system as a whole, capturing the coupled dynamics of two storages operating in tandem on the same river.

The benchmark against which these models were tested is the proposed Bekhma Reservoir, a substantially larger single storage project on the Greater Zab whose construction is to resume within Iraq. For decades, Bekhma has been envisioned as the flagship hydraulic work of the Iraqi portion of the basin, promising flood control, water storage and a significant hydropower capacity. But large dams on transboundary rivers are rarely simple engineering matters. The Greater Zab rises in Turkey before crossing into Iraq, and upstream development affects downstream flows, making any large Iraqi reservoir vulnerable to decisions taken beyond its borders. Against this backdrop, the Mosul researchers posed a question with immediate policy relevance: if two smaller reservoirs could deliver comparable hydropower, would the multi-reservoir configuration offer a more resilient and politically flexible path?

To answer it, the team ran their harmony search models under three modes of annual inflow to the reservoirs, representing different hydrological conditions. The model outputs constitute optimal policies for storage and hydropower generation for both the single-reservoir and multi-reservoir configurations. In each model run, the algorithm searched across the space of possible release and storage decisions, evaluating each candidate policy by the annual energy it would produce while respecting physical constraints on reservoir capacity, turbine throughput and water availability. The harmony memory gradually filled with the most productive operating strategies, converging toward policies that squeeze the maximum energy from every cubic meter of water passing through the turbines.

The headline result is striking. The optimal hydropower generated through the coordinated operation of the proposed two-reservoir system was found to approach the optimal power of the Bekhma reservoir, despite the two small dams holding far less water in combination. On that basis, the authors conclude that the performance of the Zab 1 and Zab 2 pair can be considered better than that of the larger single reservoir. The comparison also served as a validation exercise: the efficiency of the harmony search algorithm across both single and multi-reservoir formulations indicates that the method itself is a robust instrument for this class of problem, capable of handling the added dimensionality that coupled systems introduce.

Why should two small reservoirs nearly match one large one? The physics and economics of hydropower offer clues. Energy production scales with both the head, the height through which water falls, and the flow through the turbines. A large deep reservoir enjoys a high head, but a distributed system gains operational flexibility: releases can be timed sequentially, storages can be drawn down independently, and the cascade can adapt to variable inflows in ways a single monolithic pool cannot. Optimized coordination, the very thing the harmony search models were designed to find, is what converts that flexibility into energy. In effect, good scheduling partially substitutes for raw storage volume, a lesson with implications well beyond the Greater Zab.

The broader significance lies in what this means for water-stressed regions considering mega-dam projects. Small, distributed infrastructure can often be built faster, financed more easily, and sited with less social and environmental disruption, while an optimized multi-reservoir system may offer redundancy that a single large dam lacks. For Iraq, where the Ministry of Water Resources has repeatedly warned of diminishing flows from transboundary rivers and where hydropower remains a critical component of the electricity mix, the prospect of achieving near-Bekhma energy output from two coordinated mid-sized reservoirs reframes the planning conversation. The authors’ earlier work had already proposed this multi-reservoir system spanning the border of riparian countries and specified its operational outputs through simulation models; the present study adds an optimization layer showing the configuration can hold its own against the classic single-dam design.

None of this settles the matter, of course. The results are model-based, resting on inflow scenarios drawn from data supplied by the General Directorate of Water Resources in Erbil Governorate, and real reservoirs face sedimentation, climate-driven shifts in runoff, and the political turbulence of any river shared between nations. Harmony search, like all metaheuristics, guarantees good solutions rather than provably perfect ones. Yet the study’s core contribution stands: a music-inspired algorithm, humming its way through millions of candidate operating policies, has shown that clever coordination can narrow, and perhaps close, the gap between many small stores of water and one very large one. In a century that will be defined by how efficiently humanity manages its rivers, that is a harmony worth listening to.

Subject of Research: Optimization of single and multi-reservoir hydropower operation using the harmony search algorithm

Article Title: Modelling the Operation of Single and Multi-Reservoir Systems Using the Harmony Search Algorithm

Article References: Al-Taey, A. A., & Al-Aqeeli, Y. H. (2026). Modelling the Operation of Single and Multi-Reservoir Systems Using the Harmony Search Algorithm. Water Resources Management, 40(12), Article 529. https://doi.org/10.1007/s11269-026-04891-x

Image Credits: AI Generated

DOI: 10.1007/s11269-026-04891-x

Keywords: harmony search algorithm, hydropower, multi-reservoir system, Bekhma reservoir, Greater Zab river, water resources management, optimization model, reservoir operation, metaheuristics, Iraq, renewable energy, storage systems