Simulation-based adaption of scheduling knowledge

Nowadays, markets are changing frequently and so are the orders that were placed. Therefore, the time from ordering a product until the delivery date becomes shorter and shorter. Furthermore, production systems are subject to different exogenous and endogenous disturbances like machine breakdowns, urgent orders, material failures and so on due to companies acting in a fast and complex world. Currently available scheduling and rescheduling mechanisms are lacking in solution quality or need large calculation times. Therefore, new self-adapting systems that are able to generate good solutions quickly and refine it-self over time are needed. A new approach for a simulation based adaption mechanism for a knowledge based system is presented in this paper. Adaption of the knowledge based and the used classifier is supported by the mechanism. It is shown, that the solution quality increases when using the adaption mechanism instead of the native system without adaption component.