Research on Service Utility Evaluation Method for on-Demand Mobility Systems

In a multi-agent system, the whole social picture is redefined under the logic of service, and the system gradually turns into an ecosystem that is operated by many intelligence agents with division of labor and cooperation. As a representative of the emerging sharing economy, the on-demand mobility system presents the characteristics of multi-agent, ecology and complexity, and there are concerns such as single interest, unequal rights and responsibilities, and imbalance of agents' behaviors among platforms, drivers and passengers. In this work, we take the ondemand mobility system as the scenario, and propose a multiperspective, multi-value service utility evaluation method for multi-agent systems, which balances the behaviors of the service agents. With the simulation experiments, we apply the service utility evaluation method to the on-demand mobility system, explore the impact of the driver intelligence level and the interaction intensity on the individual utility as well as the system utility, and analyze the trend of its change to provide an optimization path for the system utility in the subsequent actual scenario.