Overlapping coordination methodology using nonfeasible decompositions

This paper focuses on computational aspects of hierarchical overlapping coordination, the concept of which was first introduced in Haimes and Macko, and proposes new coordination methodology in this paper. The overlapping coordination formulated mathematically by Macko and Haimes as the optimization methodology for large scale systems is a feasible decomposition method in the sense that every constraint is satisfied during the optimization process. Here, a new overlapping coordination method is developed which uses a non-feasible decomposition method. The Kuhn-Tucker or Lagrange multipliers for the coupling constraints of different decompositions are used as the coordination variables. The coupling constraints may not be satisfied until the optimal solution is found. Numerical examples are solved in order to show how the proposed method works and the results show that it is very promising. The proposed method, keeping the attribute of overlapping coordination concepts, can be applied effectively for many types of real large scale problems. [In Japanese.]