Abstract We consider the problem of scheduling a set of tasks whose precedence relation is representable as a directed forest, on two identical machines, in order to minimize the total completion time. A new heuristic algorithm is presented which provides a 1 4 approximate solution. This algorithm is based on the idea of critical jobs, where a job is a set of tasks corresponding to a maximal tree in the forest. A critical job is one whose total duration exceeds the total duration of all other jobs. In the paper, first a 1 4 approximate algorithm for the case in which no job is critical is presented, then this algorithm is extended to the general case. The 1 4 approximate bound is proved to be tight. The complexity of the proposed algorithm is studied.
[1]
Teofilo F. Gonzalez,et al.
A New Algorithm for Preemptive Scheduling of Trees
,
1980,
JACM.
[2]
Manfred Kunde,et al.
Nonpreemptive LP-Scheduling on Homogeneous Multiprocessor Systems
,
1981,
SIAM J. Comput..
[3]
Joseph Y.-T. Leung,et al.
Scheduling Chain-Structured Tasks to Minimize Makespan and Mean Flow Time
,
1991,
Inf. Comput..
[4]
Eugene L. Lawler,et al.
Sequencing and scheduling: algorithms and complexity
,
1989
.
[5]
Ronald L. Graham,et al.
Bounds for certain multiprocessing anomalies
,
1966
.
[6]
George L. Nemhauser,et al.
Handbooks in operations research and management science
,
1989
.