A stochastic approach to robot plan formation

Abstract

A new approach to robot plan formation is suggested and discussed, based on probabilistic ideas.The concept of incidental phenomena from [18] is extended into a randomized form and then it is used as a background for exploring approximate plans (plans with a certain degree of execution reliability and of fittnes to a given task).This makes it possible to develop plans with loops, which would otherwise require infinitistic tools.*) Making no distinction between large finite and infinite objects is characteristic for heuristic approaches in artificial intelligence.