Impact of System Design Parameters on Performance of Cooperative Agent Teams

Abstract

During the last few years several models of cooperative control among a team of agents have been proposed. In this paper, we study the impact of several system design parameters on a cooperative control methodology. An event-driven simulation model is developed to investigate how system design parameters affect the search and task performance of cooperative multi agent teams. In this study we focus on the problem of uninhabited air vehicles (UAVs) searching for targets in an uncertain environment We have previously presented a model-based decision-making and task-allocation methodology for this problem. In this paper, we investigate how the choice of two critically important parameters in the algorithm influence the performance of the UAV team. We evaluate the effect of these parameters on six different performance measures: 1) total mission time; 2) classification error rate; 3) target neutralization time; 4) task attempt times; 5) target miss ratio; and 6) task-driven traveling distance. The results provide valuable intuition on the selection of system design parameters based on different mission requirements