In a groundbreaking fusion of empirical analysis and advanced simulation modeling, researchers have unveiled new insights into how enterprises’ scientific capabilities dynamically influence their innovation performance within science-driven industries. This multifaceted study employs a novel multi-agent simulation framework, developed in the cutting-edge NetLogo environment, to capture the intricate interplay between public institutions, enterprises, and the scientific knowledge that propels technological breakthroughs. By transcending traditional secondary data limitations, this approach offers a vivid portrayal of innovation as an evolving, interactive process rather than a static outcome.
Traditional approaches to measuring scientific and technological achievements primarily rely on proxy indicators such as published papers and patent counts. While these metrics provide useful snapshots, they fall short in reflecting the dynamic relationships between enterprises’ scientific efforts and their innovation outcomes over time. Secondary data limitations—ranging from lack of timeliness to static representation—hamper deeper understanding of how enterprise-level scientific capabilities evolve and translate into technological performance. This study addresses these challenges by embedding real-world entities into a simulated ecosystem, where knowledge creation, absorption, and collaboration emerge naturally from agent interactions.
The multi-agent simulation model enables each “agent”—representing enterprises, public institutions, or abstract knowledge units—to engage autonomously according to predefined behavioral assumptions and environmental constraints. This micro-macro framework captures how hundreds or even thousands of independent actors collectively generate macro-level innovation patterns. Core assumptions guiding the simulation articulate that enterprises absorb scientific knowledge from public institutions to enhance their capacities, actively collaborate to produce new knowledge, and incur costs as well as returns during innovation activities. These assumptions ensure a realistic balance between the benefits and resource investments inherent in knowledge-based innovation.
Central to the simulation’s design is the dichotomy between tangible real entities (enterprises and public research bodies) and intangible virtual entities (scientific knowledge bases and patents). This dual-layered architecture allows researchers to probe the nuanced mechanisms whereby scientific knowledge is not only produced but also diffused and transformed into innovation output. The interaction topology among these agents—carefully parameterized in the model—reflects real-world collaborative networks, highlighting the pivotal role of joint R&D efforts in advancing industry capabilities.
The simulation unfolds in three distinct industrial development phases: gestation, formation, and maturity. During the gestation phase, public institutions dominate scientific knowledge production while enterprises play a relatively passive learning role. As the industry transitions into the formation stage, a shift toward enterprise-public institution collaboration marks increased absorption and incremental innovation. Finally, industry maturity reveals enterprises as dominant scientific contributors, underlining their evolution into knowledge leaders driving innovation performance and technological advancements at scale.
Notably, the simulation maps out the trajectory of enterprises’ internal scientific capacity growth, demonstrating how innovation performance is intricately tied to an expanding knowledge base developed both independently and through collaboration. This aligns with observed empirical phenomena where top-tier enterprises emerge as “long-tail” outliers, vastly outperforming competitors in generating scientific and technological outputs. Such stratification underscores the competitive advantages held by innovation-focused enterprises in science-based sectors.
Quantitative analysis within the simulation underscores a robust, positive correlation between an enterprise’s scientific capabilities, innovation intensity, and resultant technological performance. This interdependence confirms the hypothesis that scientific capabilities are not peripheral but central determinants of innovation success. The model thus provides compelling evidence that investments into enhancing internal scientific knowledge bases materially boost an enterprise’s capacity to innovate, leading to measurably greater returns.
Equally illuminating are the results examining the impact of varying levels of scientific research funding on knowledge output and innovation outcomes. Simulations run over extended time steps reveal that as funding increases, technological innovation performance experiences exponential growth—far outpacing the corresponding gains in scientific knowledge production. Specifically, scenarios with high funding levels showcased technological innovation peaks more than triple those of unfunded conditions, revealing a non-linear amplification effect driven by financial support.
Alongside technological gains, enterprise survival and expansion were also markedly influenced by research funding levels. Notably, higher funding correlated with larger enterprise size and more stable growth trajectories. This suggests that beyond the direct innovation benefits, scientific research funding plays a crucial strategic role in maintaining enterprise competitiveness and fostering sustainable development within science-driven industries. The implication is clear: funding science is an investment not only in knowledge but also in long-term industrial vitality.
The simulation further captures the cost dynamics inherent in knowledge creation, illustrating that enterprises encounter tangible resource consumption when pursuing scientific advancements. These costs, while significant, are justified within the model through the receptivity of returns from technological innovation, presenting an economically grounded framework linking expenditure to innovation payoff. This balance validates the strategic calculus many real-world enterprises undertake when committing to R&D activities.
By advancing a comprehensive simulation methodology that incorporates both quantitative empirical evidence and system-level dynamics, this study marks a significant step forward in innovation research. It enables researchers and policymakers to explore “what-if” scenarios, test funding strategies, and understand collaborative behaviors in silico before translating insights into real-world actions. Such tools are invaluable in designing innovation ecosystems that nurture enterprise growth and technological breakthroughs alike.
Moreover, this approach highlights the essential but evolving role of enterprises in scientific research. From initially absorbing public knowledge to eventually driving the bulk of industry research and innovation, enterprises transition into innovation hubs critical to competitive advantage. This shift also emphasizes the need for stronger, symbiotic public-private partnerships to accelerate scientific discovery and technological translation across industries.
Beyond academic theory, these findings hold profound policy implications. Governments and funding agencies can leverage insights from such multi-agent simulations to optimize resource allocation, target research incentives, and foster environments conducive to innovation proliferation. Emphasizing scientific capability enhancement within enterprises emerges as a pivotal lever in national and regional innovation strategies.
Looking forward, integrating such agent-based models with real-time data streams and machine learning could further refine predictive accuracy and responsiveness. Dynamic simulations able to adapt and self-optimize may become cornerstone tools in managing complex innovation ecosystems, enabling stakeholders to anticipate shifts, identify emerging leaders, and mitigate risks effectively.
In sum, this study drives home a compelling message: scientific capabilities within enterprises are foundational to innovation performance and, by extension, the vibrancy and competitiveness of science-based industries. Through a sophisticated blend of empirical data and cutting-edge simulation, the research charts an actionable roadmap toward fostering deeper scientific engagement and maximizing returns on innovation investments. The long-term benefits ripple across enterprises, industries, and economies, underscoring the centrality of science-driven innovation in modern industrial development.
Subject of Research: The dynamic relationship between enterprises’ scientific capabilities and their innovation performance in science-based industries, explored through empirical analysis complemented by a multi-agent simulation model.
Article Title: The impact of enterprises’ scientific capabilities on innovation performance: evidence from an empirical analysis and simulation model.
Article References:
Zhu, W., Han, W., Lu, R. et al. The impact of enterprises’ scientific capabilities on innovation performance: evidence from an empirical analysis and simulation model.
Humanit Soc Sci Commun 12, 548 (2025). https://doi.org/10.1057/s41599-025-04808-w
Image Credits: AI Generated
