Application of Ordinal Optimization for distribution system reconfiguration

This paper applies a new concept in discrete event dynamic system (DEDS) optimization, called ordinal optimization, to address distribution reconfiguration for loss minimization. Different from the conventional optimization concept to make every effort to find the best numerical solution in a deterministic approach, ordinal optimization is to efficiently find a good enough solution with an acceptable probability for a discrete event optimization problem. This paper first reviews ordinal optimization and then presents probabilistic analysis, complementary to the previous theoretic research in ordinal optimization, about the overlap of representative set and good enough set. Then, this paper applies ordinal optimization to address the distribution reconfiguration problem for loss minimization. The proposed algorithm is discussed in details. Tests on a series of sample systems are performed to illustrate the idea of ordinal optimization applied in distribution reconfiguration. The paper also analyzes and verifies the test results such as ordered performance curve (OPC) shape, solution accuracy, running time, and error distribution.