Dynamic Task Allocation Using a Pheromone-Based Approach in Factory Domain Applications

The scheduling problem in real factory manufacturing systems is comprised of number of parallel machines. Each machine is capable of processing several tasks, but it may need extra costs if the current machine state should be changed to perform a different task with the current performing task. In that case, minimizing such changes with maintaining some desired performance is recommended for maximizing the overall system performance. This paper concerns a dynamic scheduling problem and scheduling algorithm is proposed based on agent based approaches inspired from division of labor in several social insects. Depending on the stimulus of task and the corresponding threshold value, individual agent will or will not perform task. The appropriate threshold for an effective scheduling is obtained by pheromone-based approach that uses the information about processed tasks in each individual agent, and the simulation results show that the performance of the proposed approach is comparable with other conventional methods.