Robust COA planning with varying durations

COA (Course of Action) planning involves resource allocation and task scheduling. Traditionally, this problem is tackled with the assumption that task duration is constant and with the objective to minimize the makespan. In contrast to this, this paper assumes task duration can vary in a time interval and the objective is to maximize the RM (Robustness Measure) given the deadline, which makes sense to deal with the duration uncertainty. A COA planning method based on GA (Genetic Algorithm) and STN (Simple Temporal Network) is proposed and a COA planning instance is presented to illustrate the usefulness of this method.