Declarative Preferences in Reactive BDI Agents

Current agent architectures implementing the belief-desire-intention (BDI) model consider agents which respond reactively to internal and external events by selecting first-available plan. Priority between plans is hard-coded in program, so reasons why a certain plan preferred remain programmer’s mind. Recent works that attempt include explicit preferences BDI treat essentially as rationale for planning tasks be performed at run-time, thus disrupting reactive nature of agents. In this paper we propose method declarative (i.e. concerning states affairs) use them manner preserves reactivity. To achieve this, prioritization step offline, (a) generating all possible outcomes situated executions, (b) relevant subset situation/outcomes couplings representative summary each plan, (c) sorting evaluating summaries through agent’s preferences. The task several conditions translating procedural knowledge an ASP program using discrete-event calculus.