Date of Award

Spring 1997

Document Type

Dissertation

Degree Name

Doctor of Philosophy in Industrial Engineering - (Ph.D.)

Department

Industrial and Manufacturing Engineering

First Advisor

Xiuli Chao

Second Advisor

Sanchoy K. Das

Third Advisor

Layek Abdel-Malek

Fourth Advisor

Carl Wolf

Fifth Advisor

Cheickna Sylla

Sixth Advisor

Lazar Spasovic

Abstract

Assembly type job-shop scheduling is a generalization of the job-shop scheduling problem to include assembly operations. In the assembly type job-shops scheduling problem, there are n jobs which are to be processed on in workstations and each job has a due date. Each job visits one or more workstations in a predetermined route. The primary difference between this new problem and the classical job-shop problem is that two or more jobs can merge to foul' a new job at a specified workstation, that is job convergence is permitted. This feature cannot be modeled by existing job-shop techniques. In this dissertation, we develop scheduling procedures for the assembly type job-shop with the objective of minimizing total weighted tardiness. Three types of workstations are modeled: single machine, parallel machine, and batch machine. We label this new scheduling procedure as SB. The SB procedure is heuristic in nature and is derived from the shifting bottleneck concept. SB decomposes the assembly type job-shop scheduling problem into several workstation scheduling sub-problems. Various types of techniques are used in developing the scheduling heuristics for these sub-problems including the greedy method, beam search, critical path analysis, local search, and dynamic programming.

The performance of SB is validated on a set of test problems and compared with priority rules that are normally used in practice. The results show that SB outperforms the priority rules by an average of 19% - 36% for the test problems. SB is extended to solve scheduling problems with other objectives including minimizing the maximum completion time, minimizing weighted flow time and minimizing maximum weighted lateness. Comparisons with the test problems, indicate that SB outperforms the priority rules for these objectives as well.

The SB procedure and its accompanying logic is programmed into an object oriented scheduling system labeled as LEKIN. The LEKIN program includes a standard library of scheduling rules and hence can be used as a platform for the development of new scheduling heuristics. In industrial applications LEKIN allows schedulers to obtain effective machine schedules rapidly. The results from this research allow us to increase shop utilization, improve customer satisfaction, and lower work-in-process inventory without a major capital investment.

Share

COinS