BDI Agents and Constraint Logic

Using Constraint Logic Programming and a data model approach, we provide an agent with a flexible way to plan and direct its actions and to manipulate and represent its knowledge. We use constraint solving techniques developed in the KRAFT project to provide the agent with a sophisticated reasoning and deliberation process. We explore the declarative use of constraints within a BDI Agent framework to represent knowledge as complex quantified constraints and apply these techniques to a courier scenario where cooperating agents communicate, delegate and exchange desires and information using Generalised Partial Global Planning mechanisms to solve a given set of tasks.