Abstract: Optimized parallel test task scheduling is key to parallel test technology. To solve the following problems of common task scheduling algorithms: taking up much time, limited practical application area, lack of formalization analysis on resource conflict and system deadlock, timed coloured Petri-net (TCPN) is used to build the TCPN model of parallel test task scheduling. Base on the reachable marking graph of TCPN model, a novel ant colony algorithm is proposed to obtain the optimized task scheduling sequence. Multi-target optimization is used during the search process to satisfy engineering requirements, and the target function integrates total test time, instrument price, and load balance. An approach of dynamic labeling to increase the pheromone difference between feasible solutions is adopted to avoid earliness of algorithm. A simulation example is given. The practical application shows that this algorithm is valid.

Key words: parallel test, modeling task scheduling, multi-target optimization, Petri nets, ant colony algorithm