Mudança de ações em problemas de planejamento sem solução

Abstract

Automated Planning is an important area of Artificial Intelligence that deals with the process of choosing actions for an intelligent agent to achieve its goals. These actions are created by a planner, an algorithm that receives as input a planning problem, and if this problem has a solution, returns a sequence of actions (plan) that takes the agent from the initial state of the problem to a state that satisfies the goal (goal state). The planning problem does not always have a solution. In these cases there are three possible reasons: incorrect specification of the initial state, planning goal is too restrictive, actions set incorrectly or incompletely specified. In this work we deal with the third possibility, proposing a method that, through changes in the set of actions, makes the problem solvable.