Agent Collaboration in Intelligent Parking Lot Systems: Dynamic Generation of Commitment Protocol

Abstract

Agent collaboration is a fundamental part of multi-agent systems when an agent cannot accomplish a goal by itself. Such collaborations are usually regulated by commitment protocols, which are typically defined at design-time. However, in many situations a protocol may not exist or be predefined at design-time which may not fit the needs of the agents when environment changes. In order to deal with such situations, agents should be able to generate protocols at run-time. In this paper, we combined commitment with capability. Firstly, we proposed the capability matching method to generate commitment protocols dynamically at run-time. Secondly, we combined capability with commitment and extended its traditional definition. Thirdly, we compared the forms, generation time and execution time of typical and extended commitments. At last, we introduced the criteria of profit to select the optimal protocol. The application of our approach will be demonstrated in intelligent parking lot systems.