![]() | Only 14 pages are availabe for public view |
Abstract Production scheduling is the main concern of the planning department in the operational level. Many methods have been suggested to solve the different types of problems which arise in production scheduling. Thanks to the advances in modeling and simulation tools, they are becoming now very popular in many fields of applications. Petri nets is a very powerful and fast growing tool for modeling and simulation which is applied to discrete event systems. However, few researchers gave enough attention to the use of the tool in production scheduling. In this work Petri nets is applied to an important type of production scheduling systems; this is the job shop problem, besides job shops with assembly structures and flexible systems. The main objectives were to build a Petri nets model for different job shop systems, to mechanize the building of this model, to run the simulation model with five common priority rules, but tailored for the Petri nets model and to compare the results of applying different priority rules. The rules used were SPT, LPT, MWKR, LWKR, and LSR. For non-cyclic job shop problems with due dates, two models are built: forward and backward Petri nets models. In both, the same priority rules are applied. Models have been built also to simulate repetitive demand and cyclic job shop. Results of non-cyclic job shop show that, for job shop problems without due dates, the forward model is only applicable and SPT rule performs better in terms of make span. For problems with due dates satisfying a certain optimal schedule, LSR performs better than other rules in the forward model, and SPT in the backward model. Overall results show that backward model gives much better results than forward one in terms of make span and mean tardiness. Results show also that cyclic job shop gives better results than job shop with repetitive demand taking mean flow time as a measure of performance. Key Words: Petri nets, Job shop, Due dates, priority rules. |