A CONNECTIONIST PARADIGM IN THE COORDINATION AND CONTROL OF MULTIPLE SELF-INTERESTED AGENTS
Katerina Alexiou, Theodore Zamenopoulos · UCL Discovery (University College London) · 2001
Complex design and planning cases involve distributed knowledge conducted by selfinterested agents. The multiplicity of interdependencies among goals, plans and activities emerged in design and planning, directs the necessity for coordination and consensus. We can distinguish two conditions where the importance for co-ordination among the parties involved, takes different forms. These are multi-agent collaborative planning and distributed problem solving. We introduce a connectionist approach to “co-ordination via control” and we discuss the formalization of a meta-model for knowledge representation that captures interdependencies among objects emerged in complex design and planning cases. We finally discuss the development of a Web-based tool for agent control and coordination.