PSMOA: Policy Support for Data Replication

In this work, we introduce PSMOA, a novel Policy Support Multi-objective Optimization Algorithm designed for decentralized data replication. PSMOA combines the Non-dominated Sorting Genetic Algorithm III (NSGA-III) with Entropy Weighted TOPSIS to dynamically assign weights to objectives based on system policies. The algorithm focuses on optimizing replication time, cost, popularity, and load balancing. Our simulations demonstrate that PSMOA outperforms NSGA-II and NSGA-III by achieving solutions with both lower replication costs and faster replication times, all while adhering to varying policy requirements. The results illustrate PSMOA's effectiveness in optimizing data replication within complex, multi-organizational environments.