Optimal Operation of Hydropower Station by Using an Improved DE Algorithm

The mid-long term optimal operation of hydropower station is a nonlinear combinatorial optimization problem with strong constraint. This paper eliminates the problem constraints by introducing penalty function and solving this problem with DE algorithm, simultaneously, it proposes a few improvements according to the disadvantages of DE algorithm: adopt the scheme which generates differential vector to carry out mutation operation for optimum individual of each generation by randomly selecting four different individuals from population; and it adopts adaptive secondary mutation differential algorithm. The example demonstrates that DE algorithm is easy for programming realization and occupies little computer memory, and it is an effective search algorithm with rapid speed of convergence and superior precision as well as a new approach for multi-reservoir combined optimal operation to overcoming the curse of dimensionality.