Multiperiod planning for utility systems using dynamic programming

In this paper, new approach for optimal multiperiod utility planning has been proposed. To solve the nonlinear planning problem more accurately within an allowable computation time, two phase approach is introduced. At the lower level, alternatives for optimum configuration at each period are generated, and nonlinear problem is solved for each configuration. By introducing heuristics into the alternative generation step, infeasible configurations are excluded and this leads to the reduction of the computation time. At the upper level, the optimum configuration is determined using dynamic programming. More reliable and accurate solution that takes into account nonlinear characteristics of the system is obtained and the computation time is greatly reduced by the combination of nonlinear programming (NLP) and heuristics combined dynamic programming.