Optimization of parallel test task scheduling with constraint satisfaction

Parallel test task scheduling is an efficient way to shorten the final makespan of several huge test projects. Put simply, a set of test tasks should be processed on several unrelated resources, and several test tasks must satisfy the predetermined technological test order. The objective of the investigated problem is to minimize the makespan. To tackle the problem, a recursive search artificial bee colony algorithm (RS-ABC) is proposed. The recursive search procedure is developed to obtain a series of implied sequences of the predetermined technological test order on the recursive tree. The artificial bee colony (ABC) algorithm is devised to find the schedule with minimum makespan by utilizing the implied sequences. To evaluate the performance of RS-ABC, small and large size instance problems are solved, and the results are compared with those of the latest algorithm and one state-of-the-art solver. The experimental results show that RS-ABC is encouraging in solving the parallel test task scheduling problem. This work can help users design an effective test plan for the shortest completion time.