Robust Otimization Job Shop Scheduling Problem using Simulation and Genetic Algorithm
Marilda Fatima de Souza da Silva1; Luiz Lycurgo Neto1; Leila Aparecida Martins1; Fabio Henrique Pereira1
1 Universidade Nove de Julho - UNINOVE
doi:10.20906/CPS/CILAMCE2017-1268
Resumo
The problems of production sequencing have been, for more than half a century, the object of study of several researchers, in the search for the best result that can simultaneously meet several measures of performance, sometimes conflicting among themselves, such as: reduce number of exchanges and Satisfy the maximum number of customers; Increase productivity and reduce exchange time; Reduce costs and increase productive capacity, among others. This work addresses the robust optimization for the production order sequencing problem. The robustness of this optimization is characterized by the ability to remain stable even in the occurrence of variations in the arrival times of production orders. The objective is to determine and verify the validation of the sequencing of orders provided by the Genetic Algorithm (GA) when subjected to these variations. A simulation model of dynamic environment of production sequencing job shop was developed in the simulator Arena and coupled to a Genetic Algorithm. The arrival of production orders obeys a probability distribution. Three performance measures will be tested: number of orders delayed; Total time of delay and the makespan which, in this case, will be the average time of order crossing. Simulations will be performed on the following sequencing rules: Priority given by GA, FIFO (First In, First Out); LIFO (Last In and First Out); EDD (Earliest Due Data); SIPT (Shortest Imminent Processing Time); DLS (Dynamic Least Slack); LWQ (Leas Work in next Queue); CR (Critical Ratio) and LS (Least Slack). The results will be presented in a timely manner.
Palavras-chave: Robust Otimation; Genetic Algorithm; Simulation; JSSP