Case based reasoning with state transition mechanism for problem-solving in AI

A method as an enhancement to Case Based Reasoning (CBR) systems, where each solution is treated as a case and a case database is used to search for its solution. However, unlike CBR systems, every Problem Solution Procedure is treated as a transition of an initial (present) problem state, in a problem domain into a final goal state in the solution domain called as a solution state, through a path of intermediate states in the problem and solution domain. Instead of a case database, a path and pattern database is used that stores all problem and solution states and the paths to either of them including the intermediate states and the elements and methods responsible for state transition. Further on, instead of finding similarities with the case database, the path and pattern database is used to find the most feasible solution if one already exists, else a path is created using learning and the database is continuously enhanced to encompass paths to all tractable solutions, considering the present stock of resources and strategies to utilize those resources. The problem of identification and definition of problems is reduced considerably by utilizing causes and symptoms previously identified to be compared to those that have actually occurred and thus reducing the problem domain to a smaller problem space.