A Physarum-inspired algorithm for logistics optimization: From the perspective of effective distance

Abstract The logistics optimization problem has received immense attention in recent years. The existing optimization methods generally put forward distribution strategies based on physical distance or topological distance. Hence, they have inherent limitations on effectively optimizing the logistics network in real-life situations. In order to address these concerns, this paper proposes a novel optimization model based on the concept of effective distance. We first define the effective distance in logistics networks, and then implement the network optimization based on effective distance with a Physarum-inspired algorithm that overcomes the slow convergence rate of exact algorithms. The superiority of our proposed model is that suppliers can cooperate with each other to realize cost reduction, while products from different suppliers on each link remain differentiated. Numerical examples of a logistics network with multiple origin-destination pairs have shown that our proposed model (which considers both economies of scale and cooperation among suppliers in the distribution process) provides a reliable and effective cost minimization strategy. The computational performance of our proposed algorithm is also better than other algorithms such as the particle swarm optimization and genetic algorithm, as indicated in our experiments.