Bi-criteria Scheduling Problems on Parallel Machines

Mathematical programming has not been used extensively for the solution of scheduling problems. Moreover, the study of bicriteria problems on single and parallel machines is an open field for research. This thesis is aimed at developing algorithms to solve bicriteria problems more efficiently and in reasonable amount of time and with little compromise on the optimality of the solutions obtained. Two classes of problems are considered. The first class consists of scheduling unit duration tasks on parallel machines. Various combinations of primary and secondary criteria are considered and optimal seeking algorithms of polynomial time complexity are developed. The second class of problems assume general processing time tasks. An algorithm is developed for the primary criterion of total tardiness and the secondary criterion of total flow time. This algorithm is based on the solution of the underlying mathematical program and makes use of dominance relationship among the jobs and fixing of variables. Experimental results are presented regarding its performance.