Abstract
Production distribution systems are increasingly crucial because of shortenedproduct life cycles, increasing competition, and uncertainty introduced by globalization.Production distribution involves a multistage supply chain network thatconsists of factories, distribution centers, retailers, and various customers. Customerdemands fluctuate and are unpredictable, thereby causing an imprecise customer quantitydemand in each period in the production distribution model, and increasing inventoryand related costs. Most studies have addressed the production distribution problemwith certain demands or a single period. To fill the gap, this study aims to integrate the extended priority-based discrete particle swarm optimization and novel extendedpriority-based hybrid genetic algorithm for solving flexible multistage production distribution under uncertain demands in multiple periods. In particular, triangular fuzzy demands are considered for minimizing the total cost, including transportation costs, inventory costs, shortage costs, and ordering costs, in the multistage and multi-timeperiod supply chain. For validation, we designed numerical experiments to compare the proposed approaches with LINGO computational software (for small problems) and conventional genetic algorithms (for normal problems) in real settings. The experimental results demonstrated practical viability of the proposed approaches.