A Weight Vector Based Multi-Objective Optimization Algorithm With Preference And Its Application On Maximum Power Point

Recently, multi-objective optimization algorithm has been widely used in many fields, such as engineering, scientific research and so on. The main goal of these methods is to obtain a set of optimal solutions distributed in the entire Pareto Front (PF). However, in the practical application of multi-objective optimization problems, the decision makers don’t need all the optimal solutions in the entire PF, but only the solutions satisfied their preferences. To solve this problem, multi-objective optimization algorithm with preference gives rise to researchers’attention, which introduces Decision Makers’preference information designed by some strategies into traditional multi-objective methods. Thus, in this thesis, we study multi-objective optimization algorithm with preference, and propose a weight vector based multi-objective optimization algorithm with preference, which verifies the capability of searching preferences solutions. Meanwhile, we study maximum power point tracking (MPPT) algorithm on photovoltaic system, and then propose a novel MPPT algorithm.The main research work and innovations of this thesis are shown as follows:(1) This thesis proposes a weight vector based multi-objective optimization algorithm with preference. MOEA/D is a novel multi-objective optimization algorithm, which searches the Pareto front according to uniformly distributed weight vector. However, the MOEA/D doesn’t consider the Decision Makers’preference information. Based on the idea of MOEA/D algorithm, this thesis proposes a weight vector based multi-objective optimization algorithm with preference. In the proposed method, the weight vector with preference is designed, which guides the search direction to interested region of Decision Makers. Finally, the solutions around the preferred point interested by Decision Makers are found. Compared with existing algorithms in a series of test functions, the simulation results verify that the proposed method can support multiple reference points, flexibly control the extent of preferred region, have no special requirement of the position of preference points and achieve better convergence.(2) This thesis proposes a weight vector based multi-objective optimization algorithm with preference with preference for MPPT on the photovoltaic system. Photovoltaic power system only has one maximum output power point (MPP) in a particular environment. Thus, the operating point deviated MPP, the output power of the solar cell will be reduced dramatically. Therefore, MPPT is a key technology of photovoltaic power generation system. However, with the environment occurring transient or PV array locating under the partial shading, the traditional MPPT algorithm may fail. MPPT problem not only satisfies the maximum power, but also needs the minimum error, which emerges between the searching current and the model corresponding to the real current. Therefore, the MPPT problem can be transformed to preference multi-objective optimization problem so that this thesis proposed a weight vector based multi-objective optimization algorithm with preference with preference for MPPT on the photovoltaic system. Compared with the traditional MPPT algorithm, the proposed method can not need to detect the power changing in real time but only detect the change of the external environment. Because of only detecting the external environment change, it can realize to find the MPPT off-line and not lose energy caused by oscillation near the MPP. The simulation results verify that the proposed method this method has higher accuracy of MPP and the error can be controlled within the ideal range.