Evolutionary Optimization of Production Materials Workflow Processes

We present an evolutionary optimisation technique for stochastic production processes, which is able to find improved production materials workflow processes with respect to arbitrary combinations of numerical quantities associated with the production process. Working from a core fragment of the BPMN language, we employ an evolutionary algorithm where stochastic model checking is used as a fitness function to determine the degree of improvement of candidate processes derived from the original process through mutation and cross-over operations. We illustrate this technique using a case study where a baked goods company seeks to improve production time while simultaneously minimising the cost and use of resources.