Searching in a multidimensional space

The generalization of a multiple-rule A* algorithm (MRA*), which provides a framework in which many evaluation functions can be combined into a multiresolutional hierarchy, is described. This algorithm was originally applied by the authors to the problem of path planning in two-dimensional space for an autonomous vehicle. The generalization of the rules that guide this algorithm allows it now to be applied to the problem of determining optimal paths through n-dimensional space. The results of applying the algorithm to the search for an optimal path in three-dimensional space are reported. Techniques for reducing the computational complexity of the algorithm are also described.<
>