Spike response threshold model for task allocation in multi-agent systems

Abstract

This paper focuses on task allocation problem in multi-agent system and we consider the solution for regulating the proportion of groups performing tasks equally to the proportion of task demands. The response threshold model inspired by division of labor observed in insect societies is applied. Each agent converts the information from surrounding environment to the spike and decides its task based on the response threshold model. The decision of an agent has partial control over its environment, but the overall system shows a desired performance. The algorithm is implemented using simulated robots and demonstrates its adaptivity with changes for task demands and the number of agents in a group.