Modelling of parallel production system with rework paths and its GA based simulator for optimal design

Production lines are widely used in high volume industries and vary in their sophistication from simple to the complicated structured such as parallel, reworks, feed-forward, etc. One of the common production styles in many modern industries is the parallel production system with rework path (PPS-RP) and one of the methods used for studying the PPS-RP design is through genetic algorithm (GA). As a one of the important tasks in using GA is how to express a chromosome. This paper attempts to find the nearest optimal design of a PPS-RP that will maximise production efficiency by optimising the following two decision variables: buffer size between each pair of work stations and machine numbers in each of the work stations. In order to do this, a new GA-simulation based method to find the nearest optimal design for the proposed PPS-RP is introduced. For efficient use of GA, the used GA methodology is based on a technique that is called non-homogeneous gene arrangement method (NGAM) which arranges the genes inside individuals. An experimental numerical examples showed that after a number of operations based on the proposed simulator, it was possible to get the nearest optimal design of PPS-RP.