Modeling Academia–Industry Collaboration in India Using Evolutionary Game Theory and Reinforcement Learning

S, Theajeshvar (2025) Modeling Academia–Industry Collaboration in India Using Evolutionary Game Theory and Reinforcement Learning. Masters thesis, Indian Institute of Science Education and Research Kolkata.

Text (MS Dissertation of Theajeshvar S (20MS157)) 20MS157_Thesis_file.pdf - Submitted Version Restricted to Repository staff only Download (921kB)

Abstract

India is at a critical juncture in its journey to become a knowledge-driven economy. Despite a series of ambitious initiatives aimed at enhancing innovation and technology commercialization, academia-industry collaboration remains sporadic and underutilized. This thesis examines the strategic misalignments and incentive challenges that underlie this phenomenon, using a game-theoretic simulation-based approach. By integrating the frameworks of evolutionary game theory and reinforcement learning with narrative economics, we build and test a dynamic agent-based model that captures how institutions adapt their collaboration strategies over time. The model consists of three comparative setups: a replicator dynamics model, a Boltzmann softmax model, and a full Q-learningbased adaptive model. These are used to simulate strategic interactions between academic and industrial agents under varying incentive structures, reputational risks, and trust environments. The simulations demonstrate that trust, repeat interactions, and public narratives play an outsized role in determining equilibrium collaboration strategies. To contextualize the findings, the thesis presents a comparative, data-driven case study of leading Indian academic-industry partnerships—most notably the IIT Madras–Siemens Centre of Excellence—and compares their structures and outcomes to national trends. Drawing on government funding records, publication statistics, and patent data, the analysis reveals that most collaborations succeed not just due to financial incentives, but due to well-structured institutional frameworks and a history of repeated interaction. Finally, we draw broader policy implications for the Indian innovation ecosystem and recommend the inclusion of the use of phased funding mechanisms, enhanced narrative framing by the government, and greater institutional support for early trust-building.

Actions (login required)

View Item