Centralised Versus Market-Based Control Under Environment Uncertainty: Case of the Mobile Task Allocation Problem (MTAP)

Abstract

This paper aims at comparing the centralised versus the market-based approach. This is done in the context of the mobile task allocation problem (MTAP) from the perspective of environmental uncertainty. MTAP is defined as an optimization problem for planning the assignment of service tasks to mobile workers. Environmental uncertainty is introduced through the injection of stochastic tasks and dynamic travel delays. A multi-agent simulator is employed to experiment the behaviour of each approach in reaction to different uncertainty levels. Preliminary results suggest a tentative conceptual model to evaluate the suitability of each approach to address MTAP in function of uncertainty. It is suggested that uncertainty’s effect on achieved performance is moderated by the timeliness of decision making, workers’ degree of local knowledge, and problem’s complexity and size.