Research On Multiobjective Evolutionary Algorithms And The Application In Load Dispatch Problems

With the implementation of a series of electricity reforms, the thermal companies need to consider reducing the operation cost of multiple perspectives. Therefore, it is important to implement optimal load dispatch among the units to improve the operational efficiency and reduce the production cost. Meanwhile, with the development of the electrical market, the load dispatch should also satisfy the adjustment time. As people increasingly attach importance to environmental protection, the harmful gases emitted from the power plant such as NOx, SO2 should also be minimized. So it is meaningful to adapt the multi-objective optimization strategies to solve the power plant load dispatch problems.In this paper, a real coded chaos genetic algorithm is proposed to solve the economic load dispatch problem. The method improves the main operators of simple genetic algorithm. And the best individual of each generation is optimized with mutative chaos optimization strategy. The proposed method overcomes the shortcoming of simple genetic algorithm to some extent, such as easy to early maturity and local convergence.As solving the multi-objective load dispatch problems, the multi-objective evolutionary strategies based on the concept of Pareto optimal have obvious advantages compared with the traditional methods, such as weigh sum method. In this paper, two multi-agent multi-objective evolutionary algorithms are proposed and applied to the multi-objective load dispatch problem. Firstly, a real-coded multi-agent multi-objective evolutionary algorithm is proposed. In this algorithm, new operators for multi-objective problem are designed, by which the agents in neighborhood interact with each other and generate the non-dominated solutions unceasingly. Meanwhile, parts of the non-dominated solutions in archive set are optimized again to assure the distribution. Secondly, a quantum multi-agent multi-objective evolutionary algorithm is proposed, which integrated quantum theory and multi-agent technology. This method can produce a large number of non-dominated solutions in iterative process. To assure the distribution and the efficiency, the adaptive grid strategy is applied and the most representative solutions are kept. The simulation results show the effectiveness of the two methods.The main innovations of this paper are as fellows:1.A real coded chaos genetic algorithm is proposed and applied to economic load dispatch problem; 2.Applied the classic algorithm NSGA-â…¡to multi-objective load dispatch problems, and compared with weigh sum method;3.Proposed a real coded multi-agent multi-objective evolutionary algorithm and applied it to multi-objective load dispatch problems;4.Proposed a quantum multi-agent multi-objective evolutionary algorithm and applied it to multi-objective load dispatch problems;...