Efficient temporal propagation algorithm for scheduling

Scheduling plays a major role in improving the productivity of many workplaces by minimizing the overall turnaround time and maximizing the resources. An iterative scheduler starts with an unfeasible schedule that satisfies only the temporal constraints and then it iteratively repairs other constraint violations, such as resource or state. After repairing each resource or state violation, the network is checked for temporal constraint satisfaction; violation of temporal constraints are repaired if possible. It is clear that the efficiency of temporal constraint propagation will significantly influence the overall efficiency of iterative repair scheduling. Many scheduling systems use either the Floyd-Warshall algorithm or Bellman-Ford algorithm for propagating metric temporal constraints. Each of these algorithms terminates in O(n/sup 3/) time that is very expensive if the algorithm is invoked several times before a feasible schedule is found. The authors propose an efficient algorithm for propagating metric temporal constraints which runs in time bounded by O(n) to O(n/sup 2/).