Coordinating mobile activities by integrating simulated and physical software agents

This study addresses a mobile user coordination problem (MUCP), in which mobile users can change their goals by scheduling or re-scheduling their tasks via mobile devices. Due to service capacity and unforeseen events, conflicts of schedule may occur among mobile users. In this study, we formulated the MUCP as a Multi-objective Context-dependent Distributed Constraint Optimization Problem (MCDCOP). To solve the MCDCOP, we developed a novel Two Stage Distributed Optimality Reaching Approach (2S-DORA), and implemented it using an integrated multi-agent system called Simulated and Physical Agents (SPA). SPA can facilitate the cooperation between mobile users in resolving schedule conflicts to maintain approximately joint optimality. We then took a traveling backpacker problem as an example and then conducted four experiments to illustrate and evaluate 2S-DORA's performance. The results show that SPA and 2S-DORA effectively solved the MCDCOP.