On the analytic representation of the Leximin ordering and its application to flexible constraint propagation

We discuss the basic formulation of constraint propagation problems and extend it to the flexible constraint propagation environment where constraints are represented as fuzzy subsets. Some methods for ordering alternative solutions with respect to a collection of flexible constraints are discussed along with their drawbacks. Among the methods introduced is the Leximin method where we note its lack of an analytic formulation. With the aid of the ordered weighted averaging operator we suggest an analytic formulation for the Leximin method. Some properties of this formulation are provided. We then describe the application of this new formulation for the Leximin method to situations in which the constraints are describable in a linear fashion. We show how we can use mixed integer programming techniques to find an optimal solution.