| Summary: | Over many decades, a large number of complex optimization problems have brought researchers' attention to consider in-depth research on optimization. Production scheduling problem is one of the optimization problems that has been the focus of researchers since the 60s. The main problem in production scheduling is to allocate the machines to perform the tasks. Job Shop Scheduling Problem (JSSP) and Flexible Job Shop Scheduling Problem (FJSSP) are two of the areas in production scheduling problems for these machines. One of the main objectives in solving JSSP and FJSSP is to obtain the best solution with minimum total completion processing time. Thus, this thesis developed algorithms for single and hybrid methods to solve JSSP and FJSSP in static and dynamic environments. In a static environment, no change is needed for the produced solution but changes to the solution are needed. On the other hand, in a dynamic environment, there are many real time events such as random arrival of jobs or machine breakdown requiring solutions. To solve these problems for static and dynamic environments, the single and hybrid methods were introduced. Single method utilizes Artificial Immune System (AIS), whereas AIS and Variable Neighbourhood Descent (VND) are used in the hybrid method. Clonal Selection Principle (CSP) algorithm in the AIS was used in the proposed single and hybrid methods. In addition, to evaluate the significance of the proposed methods, experiments and One-Way ANOVA tests were conducted. The findings showed that the hybrid method was proven to give better performance compared to single method in producing optimized solution and reduced solution generating time. The main contribution of this thesis is the development of an algorithm used in the single and hybrid methods to solve JSSP and FJSSP in static and dynamic environment.
|
|---|