Metaheuristics for Periodic Vehicle Routing with Consistency and Synchronization Constraints in Nanostore Sales and Delivery
摘要
This study introduces a Consistent Synchronized Periodic Vehicle Routing Problem (ConSyncPVRP), an extension of the Periodic Vehicle Routing Problem (PVRP) that integrates presellers and delivery trucks to serve nanostores in emerging markets. ConSyncPVRP ensures consistency in preseller-client relationships while optimizing deliveries. A mixed-integer linear programming model extends the PVRP with synchronization and consistency constraints to solve small ConSyncPVRP instances. To solve larger instances, hybrid metaheuristics are developed with greedy construction, solution repair, random destruction, and Variable Neighborhood Descent (VND) operators. Three metaheuristics—MS-ELS, MS-ILS, and GRASP—are tested on adapted benchmarks. Results show that MS-ILS achieves superior performance in large-scale instances, highlighting the advantages of iterative solution construction and improvement to find feasible and high-quality solutions when complex vehicle routing problems like the ConSyncPVRP are tackled.