A Community-Affinity-Based Evolutionary Game Approach for Resource Allocation in Industrial Networks of Service

Industrial service networks, characterized by dynamicity, heterogeneity, and nonlinearity, exhibit complex dependency relationships and interaction patterns among nodes. Based on static optimization or heuristic algorithms, traditional resource allocation methods fail to address these dynamic characteristics and the intricate node relationships, particularly in scenarios where competition and synergy coexist. This paper uses Smith dynamics to investigate the resource allocation problem within complex industrial service networks. This article proposes a novel community-affinity-based approach that integrates both network topology and node relationships. The networks are modeled as graphs, with nodes representing workstations and edges facilitating resource circulation. By introducing an affinity matrix, the study captures the varying levels of competition and synergy in resource exchange between nodes. The problem is formulated as a constrained optimization task to maximize the network's total profit. Through extensive experiments and theoretical analysis, we demonstrate that our approach effectively achieves optimal resource allocation under various network structures. The results show that our method successfully drives the system to a stable and optimal state while accounting for topological influences, providing valuable insights for practical industrial applications.