abstract
- Manufacturing companies are being pushed to their limits due to an increase of production complexity guided by a growing standards demand by the costumers. To respond properly to this, manufacturing companies must adopt innovative control architectures that are able to handle better the occurrence of disturbances at shop-floor level (e.g. workstation breakdown, orders cancellation or modification). Additionally, the selection of a proper scheduling algorithms assumes a crucial point, in the sense that the increase of optimization levels depend on this. This paper presents a Genetic Algorithm (GA) based technique to be embedded into the supervisor entity present at the ADACOR2 aiming to improve the existing fast and non-optimal scheduling technique, improving the overall system processing execution. The main requirements of the GA is to be fast enough to be usable in demanding environments improving the optimization output. The proposed algorithm is tested using a Flexible Manufacturing System using different configurations of transportation and batch sizes. Results show that despite the presented GA technique increased the optimization calculation time it performs better considering the sum of this time with the gain in the optimization output.