A non-deterministic planner for planetary rovers

This paper presents an integrated planning and scheduling algorithm based on co-evolutionary algorithms. This planner is the core of the delibera-tive level of a three-layer autonomous system (called Wisdom) for planetary rovers. The planner operates at two different levels: at a higher level reallocates and transforms mission goals, based on contingent events, in order to reach scientifically interesting targets and minimize the risk of a failure, at a lower level it generates scheduled sequences of actions that optimize a number of objectives. A particular implementation of co-evolutionary algorithms is used to generate sets of Pareto-optimal plans for every given sequence of goals. Some tests will illustrate the main characteristics of the algorithm