Ant Colony Optimization Algorithms to Enable Dynamic Milkrun Logistics

Flexibility in combination with high capacities are the main reasons for milkruns being one of the most popular intralogistics solutions. In most cases they are only used for static routes to always deliver the same material to the same stations. However, in the context of Industry 4.0, milkrun logistic also has become very popular for use cases where different materials have to be delivered to different stations in little time, so routes cannot be planned in advance anymore. As loading and unloading the milkrun requires a significant amount of time, beside the routing problem itself, both driving and loading times have to be taken into account. Especially in scenarios where high flexibility is required those times will vary significantly and thus are a crucial factor for obtaining the optimal solution. Although containing stochastic components, those times can be predicted by considering the optimal point of time for delivery. In consequence, the best tradeoff between short routes and optimal delivery times is in favor of the shortest route. To solve this optimization problem a biology-inspired method - the ant colony optimization algorithm - has been enhanced to obtain the best solution regarding the above-mentioned aspects. A manufacturing scenario was used to prove the ability of the algorithm in real world problems. It also demonstrates the ability to adapt to changes in manufacturing systems very quickly by dynamically modelling and simulating the processes in intralogistics. The paper describes the ant colony optimization algorithm with the necessary extensions to enable it for milkrun logistic problems. Additionally the implemented software environment to apply the algorithm in practice is explained. (C) 2017 The Authors. Published by Elsevier B.V.