Optimising supply chain networks by means of a hybridised simulation-based approach

Supply chain management, in the scope of the described project, is about managing the flow of materials in a network of factories producing and transforming raw material into intermediate and final product, and the use of buffering instances such as storage tanks, silos and stockpiles. The system we present tries to reconcile the drivers of a supply chain: demand for final product and supply of raw material. In addition to balancing the material flow to honour physical constraints (i.e. storage capacities, minimum production rates, transportation bottlenecks, etc.), the system aims to maximise the overall output of the supply chain network. Other benefits from a business point of view are the reduction of time to generate a factory plan while providing better accuracy and visibility of the material flow. Reducing the costs for creating a plan allows for what-if-scenario analysis and strategic planning which would not have been possible otherwise. In order to optimise the material flow, an Evolutionary Algorithm (EA) was employed that incorporates operators handling business and general planning constraints. Furthermore, the EA utilises a discrete-event simulation (DES) with characteristics of continuous simulations as part of its fitness evaluation. We present preliminary results obtained from a project carried out in cooperation with an Australian ASX listed company manufacturing agricultural chemicals.