Meta-heuristic Method Of Energy Units Scheduling And System

Energy units scheduling is an important class of typical combinatorial optimization problems. The problem is to determine the start-up and shut-down schedule of energy units to meet system demand so that the total production cost is minimized while satisfying constraints. It’s important to make energy units scheduling decision to improve economic efficiency and reduce system energy consumption, etc. Since most of production scheduling problems have been proved to be NP-hard, this thesis proposes particle swarm optimization for solving the problem. The content is summarized as follows:(1) According to the practice production process, a nonlinear mixed integer programming model with the unit open constraint is formulated. The objective is to minimize the system cost. Since the model cannot be solved using the commercial software, the nonlinear constraints are transformed into linear constraints, and then computational experiments demonstrate that the proposed model are very efficient.(2) Since the commercial software cannot solve the large-scale instances within a limit time, this thesis proposes a binary particle swarm optimization with a heuristic strategy (BPSO) to solve energy unit scheduling problem. A heuristic strategy is designed to adjust the particles satisfying all the constraints and obtains the fitness value of each particle.(3) For solving energy units scheduling problem, BPSO optimization has shortcomings, such as low convergence speed and solution precision. This thesis proposes the continuous decoding method to improve particle swarm optimization (IPSO). The method is relaxing the discrete variables, mapping the discrete variables to the continuous space coding, and then decoding particle into unit commitment status. This algorithm is tested based on randomly generated instances. The results show that this algorithm converges fast and has high solution precision by comparing experimental results with BPSO and the commercial software.(4) Based on the model and particle swarm optimization algorithm, this thesis designs the decision support system framework and functional modules of energy unit scheduling. Theoretical research results are embedded in the decision support system.