BACKGROUND AND OBJECTIVES: The management of high-level nuclear waste is a pressing global challenge, with over 400,000 metric tons in temporary storage worldwide. International disputes frequently arise due to disagreements over site selection, cost allocation, environmental risks, and long-term liability, often leading to negotiation deadlocks. Existing governance frameworks lack structured mechanisms to balance the competing objectives of multiple stakeholders, including waste-producing ("Disposer") nations and potentially affected ("Affected") nations. This paper aims to resolve these transboundary nuclear waste treatment conflicts by developing a hybrid analytical model that integrates Game Theory to model strategic interactions and the ε-Multi-Objective Evolutionary Algorithm (ε-MOEA) to find optimal solutions that balance competing goals such as cost, environmental safety, and economic benefits.
METHODS: This manuscript employs a combined Game Theory and computational algorithm approach to resolve international nuclear waste disputes. Game Theory models the conflict between two player types: disposer countries (waste producers) and affected countries (those impacted by disposal). Each player has specific goals, strategies, and costs, with the model seeking a Nash Equilibrium-a stable agreement where no country can unilaterally improve its outcome. Due to the problem's high complexity with multiple competing objectives, the study utilizes the ε-MOEA optimization algorithm. This algorithm efficiently explores millions of possible strategy combinations to identify optimal compromises, balancing outcomes to ensure fair and practical solutions for all involved countries.
FINDINGS: Computational experiments compared ε-MOEA against other multi-objective algorithms (NSGA-II, NSGA-III, PESA2, VEGA). The key finding was that all algorithms converged to the same optimal fitness value (-5280.33), demonstrating the model's robustness in identifying a stable equilibrium. However, ε-MOEA achieved this result with the shortest and most stable runtime (approximately 5.00 seconds per iteration), significantly outperforming other algorithms in computational efficiency. This indicates that ε-MOEA is particularly well-suited for solving this complex, high-dimensional problem efficiently, providing a diverse set of Pareto-optimal solutions for policymakers to evaluate trade-offs.
CONCLUSION: A hybrid Game theory and ε-MOEA framework effectively resolves international nuclear waste conflicts by modeling strategic interactions and optimizing for multiple objectives. This scalable approach identifies stable, fair agreements that balance the interests of both producing and affected nations, supporting sustainable international governance. Future work should focus on improving computational efficiency for larger numbers of players.