Optimal and suboptimal decisions in perceptual problem-solving tasks
By Amnon Rapoport · Systems Research and Behavioral Science · 1969
Certain perceptual problem-solving tasks can be formulated as multistage decision-making problems. The dynamic programming approach can then be used to find the optimal decisions. When constraints are imposed on the problem-solver's ability to scan his environment and to process information, simple algorithms for solving the problem can be constructed. The algorithms yield suboptimal solutions that may adequately describe the problem-solver's behavior. Elithorn's perceptual maze test is reanalyzed and the results of an illustrative experiment are presented and discussed.