Optimizing the coordinated inventory and routing model with joint replenishment and location under (R, S) Policy and non-stationary stochastic demands

Abstract

In practice, more operations are integrated into supply chain management to meet increasing market competition. This study is the first to examine an extensive, coordinated joint replenishment, inventory, location, and routing problem under multi-period, non-stationary stochastic demands. To model this problem, a series of intervals covering the replenishment cycle over the planning horizon is employed, significantly simplifying the calculation of average inventory holding and shortage costs under three dynamic factors: dynamic safety stock, review policy (Rt, St), and varying lead time. A novel two-phase heuristic algorithm is proposed to solve the problem. In the first phase, a self-adaptive differential evolution algorithm with random neighborhood search (SDE-RNS), proven superior across experimental comparisons with eight other algorithms, solves the coordinated joint replenishment, inventory, and location problem using one-to-one direct dispatching. This step generates an initial solution for the subsequent phase. In the second phase, the tabu search algorithm, featuring random selection among multiple Moves, optimizes the routing schedule and coordinates the integrated joint replenishment, inventory, location, and routing decisions. The efficiency of this random move approach within the tabu search is validated through comparative experiments against an individual move strategy. The overall utility of the first phase is demonstrated by comparing the performance of the two-phase algorithm with that of using only the second phase. Finally, a sensitive analysis of the parameters provides valuable managerial insights by revealing their distinct effects on overall costs, replenishment frequency, and the number of distribution centers.