Solving a backtracking intelligently : a case study / Muthukkaruppan Annamalai

This paper studies a special kind of Constraint Satisfaction Problem (CSP) related to a case of resource allocation. It attempts to study the procedure of allocating a collection of resources to a group of contenders, in a satisfactory manner. In this study, we have modeled the CSP as a constraint graph. This allows us to device an augmented backtracking algorithm that could be employed to find a solution for this CSP. There are three important features of this algorithm that we would like to highlight here, namely I) minimally backtracking to resolve the constraint violation, II) making use of the information associated with a failed search to prune the search space and III) not repeating a computation. This algorithm was tested on the New Vehicle Sales System application (Mohd Zamri, 1998) and has been found to consistently produce the desired result.