Particle Swarm Optimization for Tackling Continuous Review Inventory Models

We propose an alternative algorithm for solving continuous review inventory model problems for deteriorating items over a finite horizon. Our interest focuses on the case of time–dependent demand and backlogging rates, limited or infinite warehouse capacity and taking into account the time value of money. The algorithm is based on Particle Swarm Optimization and it is capable of computing the number of replenishment cycles as well as the corresponding shortage and replenishment instances concurrently, thereby alleviating the heavy computational burden posed by the analytical solution of the problem through the Kuhn–Tucker approach. The proposed technique does not require any gradient information but cost function values solely, while a penalty function is employed to address the cases of limited warehouse capacity. Experiments are conducted on models proposed in the relative literature, justifying the usefulness of the algorithm.