![]() | Only 14 pages are availabe for public view |
Abstract This thesis extends the research on rough mathematical programming.Firstly, we present a generalized classification of the rough mathematical programming problems depending not only on the place of roughness in the problem but also on the granularity level of the search space. We show that the optimal solutions of a rough mathematical programming problem could be simple points from the fine universe or equivalence classes from the coarse universe. Rough set theory and granular computing have been used in modeling the classes and characterizing their optimal sets. Secondly, we present a new kind of problems called ”uncertain rough mathematical programming problem” which induced by data loss. We focus on problems that have a single crisp objective function and a rough feasible set whose approximations are experiencing a lack of the available information. Finally, we introduce how to apply arithmetic operations on rough functions. |