Flexible machine allocation and buffer optimization in a serial production lines

Abstract

In this thesis, we propose an optimization model to achieve the maximum production rate by allocating the machine sequence as well as the buffer sizes in the line. Manufacturing industry such as LCD Panel Glass Manufacturing or Car Paint assembly line utilizes intermediate buffers between to maintain the production continuity and to achieve maximum production rate. In practical sense buffer space is finite and limited within factory shop and research to optimize buffer size has been carried on for years in order to get the maximum production rate within the available space. The system in this research is assumed as a deterministic processing time of a serial production line with geometrically distributed MTTF (Mean Time to Failure) and MTTR (Mean Time To Repair). The machines have equal processing time and are subject to failure. Furthermore, inspired by the aforementioned manufacturing industry, we propose new optimization model that is able to improve the production rate by changing the sequence of flexible machines. To ensure the buffer size optimality we use Heuristic approach called Sectioning approach in evaluating every sequence. Machine Allocation analysis is started with Full Enumeration with k! of possible solution which search space is increasing exponentially. To evaluate longer lines we propose the use of genetic algorithm and particle swarm optimization. In this research we present how the algorithm works followed by the result of both methods which resulted in smaller search space and reduction of computation time.