Stackelberg-game-based energy optimization strategy for interactive energy use of flexible manufacturing industrial parks

With the emergence of distributed energy resources coexisting with the power grid, optimizing energy utilization across the entire industrial park while balancing the interests of park operators and manufacturing plants has become increasingly critical. Existing optimization scheduling strategies often fail to capture the complexity of flexible production processes involving multiple production lines, multiple operations, and multiple products, or lack the capability to implement Stackelberg-game energy optimization at the “park-level” with multiple manufacturing plants. To address these limitations, this paper proposes an energy optimization strategy for flexible manufacturing industrial parks based on the Stackelberg-game framework. The energy optimization models for the industrial park (“leader” problem) and the flexible manufacturing plants (FMPs, “follower” problem) are developed, and the Stackelberg-game optimization problem is reformulated into a three-layer nested optimization framework, consisting of outer, middle, and inner layers. The proposed framework is solved using a Bayesian Optimization method with Nested Mixed Inter Linear Programming (BO-NMILP). The case study results indicate that the proposed energy optimization strategy for flexible manufacturing industrial parks, based on the Stackelberg-game framework, can enhance the self-consumption of photovoltaic power within the park, maximize the industrial park's revenue, and effectively handle the complex multi-line, multi-operation, and multi-product production processes of FMPs to meet production demands. Additionally, it significantly reduces the electricity costs of FMPs while improving overall energy efficiency. • Stackelberg-game-based energy optimization for flexible manufacturing industrial parks. • The “follower” problem considering flexible production processes. • Flexible production process considering multiple machines in parallel. • Bayesian optimization method with nested mixed integer linear programming.