The problem of scheduling n independent jobs on m parallel processors with the objective of minimizing makespan has been studied by several researchers. The non-premptive version of this problem is known to be NP-complete. Several efficient heuristics have been suggested for this problem with equal, uniform, and unequal processors. In this paper the problem is considered with the objective of minimizing some measures of job tardiness. Three heuristics are suggested and their computational behaviors are studied. A statistical analysis has also been conducted to compare the effectiveness of the heuristics.
