Solving Problems in an Uncertain World
Stuart M. Shieber · Digital Access to Scholarship at Harvard (DASH) (Harvard University) · 2013
The issue of problem solving in the context of incomplete or inconsistent information is a precursor to designing multiple agent planning systems. We present some general principles of planning and problem-solving in uncertainty, and instantiate these principles in a problem-solving system based on the NOAH planning system. The new NOAH system, working on the blocks world as a test domain, plans in certainty with the same efficacy as the original system, but can also handle a large class of errors caused by inconsistent or incomplete information.