A solver for inconsistent continuous satisfaction problems: An application to automobile shock-quality testing
In this thesis, we are primarily interested in Constraint Programming where the problem is defined over continuous domains. ^ In particular, we are concerned about two opposite situations that can surface during Constraint Solving, a particular branch of Constraint Programming. Ideally, you would like to determine a unique solution, but because this is not always the case, our research focuses on cases when no distinct solution exists and too many solutions are found. In the case where no solution is found, the model of the problem is said to be inconsistent or incompatible. However, sometimes a solution is necessary and simply stating that no solution exists is not acceptable. As a result, some form of flexibility or relaxation of the original problem can be introduced to the solving process so a weaker solution set can still be reached. Therefore, studies on this flexibility sparked a research area known as flexible constraints. In the latter case, one is to wonder which solution among many is best. Thus, finding the optimal or best solution among the entire solution set can be thought of as a constrained optimization problem. Intuitively, the idea is to filter the solution set by adding a criterion over the entire solution set to produce a more suitable solution. ^ Consequently, the center of our work focuses on both optimization problems and the development of tools used to model flexible constraints. (Abstract shortened by UMI.) ^
Engineering, Automotive|Computer Science
Coy, Richard Garrett, "A solver for inconsistent continuous satisfaction problems: An application to automobile shock-quality testing" (2005). ETD Collection for University of Texas, El Paso. AAI1430239.