Maximizing the Percentage of On-Time Jobs with Sequence Dependent Deteriorating Process Times

The paper addresses the problem of maximizing the percentage of on-time jobs in a parallel machine environment with sequence dependent deterioration. The deterioration of each machine (and therefore of the job processing times) is a function of the sequence of jobs that have been processed by the machine. Two machine loading strategies are combined with a set of list scheduling algorithms to solve the identical and unrelated machine versions of the problem. The proposed solutions approaches are tested using a large set of problem instances that consider various levels of the number of jobs and machines, the due date tightness, and the deterioration effect. The results indicate that the approach based on loading considering all machines simultaneously and assigns jobs by due date is the most effective.

Static vs. Dynamic List-Scheduling Performance Comparison

The problem of efficient task scheduling is one of the most important and most difficult issues in homogeneous computing environments. Finding an optimal solution for a scheduling problem is NP-complete. Therefore, it is necessary to have heuristics to find a reasonably good schedule rather than evaluate all possible schedules. List-scheduling is generally accepted as an attractive approach, since it pairs low complexity with good results. List-scheduling algorithms schedule tasks in order of priority. This priority can be computed either statically (before scheduling) or dynamically (during scheduling). This paper presents the characteristics of the two main static and the two main dynamic list-scheduling algorithms. It also compares their performance in dealing with random generated graphs with various characteristics.