Flexible job shop scheduling of a production line with overlapping in operations

Abstract

Creating a schedule is of utmost importance for manufacturing industries so that a company can have an accurate plan of the time frames at which a number of goods will be produced. Prior planning allows company directors to plan ahead, and order enough stock to produce all the required goods. Different studies focus on identifying different approaches that can be used to perform this process, and to analyse the performance of the approaches implemented. Since an order consists of processing a number of operations, the process of scheduling involves, the allocation of each operation to a specific machine that can process such an operation, and the sequencing of these operations to determine the order at which the operations are processed on their respective machines. A system was developed for the purpose of this study, in order to conduct experiments on a number of approaches that were developed, and to carry out an assessment and comparison of their performance. The approaches implemented were those of the Ant Colony Optimization (ACO), Genetic Algorithm (GA), and Tabu Search (TS), with the method of assessment being the makespan, that is, the total time taken to process all the operations pertaining to a set of jobs. In this dissertation, two types of data sets are assessed. The first data set is a real data set provided from a manufacturing industry. The data given consists of information about the machines available and stock required to process each operation, a set of orders for which a number of goods would be produced, and a corresponding schedule created manually by a company employee. Another type of data set to be introduced consists of a number of well known problems. This data set is used to compare the performance of the approaches implemented, against those of well known problems in order to assess whether the approaches implemented are better than those of other studies, given problems of different sizes and complexity.