Research On Multi-objective Particle Swarm Optimization Method For Building Energy-Saving Design

Building energy efficiency is an important issue in the field of energy consumption.Improving building energy efficiency has become an international issue for designers and researchers to consider,which poses a greater challenge to academia and industry.Building energy-saving problems often need to use energy consumption simulation softwares to complete the whole process.Because of the complexity of building energysaving problems,more new algorithms and tools are needed in order to quickly find the optimal solution of building model problems.Particle swarm optimization(PSO)has been widely used in various practical problems because of its simple concept and fast convergence speed.However,it is undeniable that the traditional particle swarm optimization algorithm needs to adjust the inertia weight and learning factor according to the actual problems,so as to optimize the global search and local exploitation ability of the algorithm.This method has the disadvantage of being sensitive to the values of inertia weight and learning factor.How to improve the global and local search ability of the algorithm,and how to save the time of the building energy-saving optimization,still are open problems that need to be solved.On the basis of particle swarm optimization,this thesis studies the problem of building energy conservation from the algorithm level.The main research work of this paper is as follows:(1)In view of the sensitivity of particle swarm optimization to the values of the inertia weight and learning factor,an improved and concise multi-objective particle swarm optimization algorithm based on adaptive disturbance factor is proposed.Based on the bare-bones multi-objective particle swarm optimization algorithm,a disturbance factor determined by the global and individual extreme points is introduced,which effectively improves the global optimization ability of the algorithm.The proposed multiobjective optimization algorithm is used to deal with the buildings with single-room and multiple-rooms.The experimental results show that this propsoed method significantly improves the convergence and distribution of Pareto optimal solution set,and is a high competitive optimization method to solve the problem of building energy saving design.(2)Aiming at the time-consuming problem of exsiting building energy saving optimization methods,a surrogate-model-assisted multi-objective particle swarm optimization method for building energy saving design is proposed.This method determines the update timing of the surrogate-model and the size of samples used in the evolution process to achieve an effective balance between the accuracy of the model and the cost of model management.This method presents a variable-sample-size surrogate-model management strategy guided by dual archivesets.On the basis of guaranteeing the accuracy of the surrogate-model,this strategy can reasonably reduce the number of individuals to be truely evaluated,thereby reducing the update cost of the model.Next,a decomposition-based surrogate-model assisted multi-objective particle swarm optimization algorithm is presented.Under the decomposition framework of MOEA/D,new operators,such as the proposed surrogate-model management strategy,the few-parameter bare-bones particle update strategy,and the crowding-based population initialization strategy,are used to effectively improve the performance of the proposed algorithm.Applied the proposed algorithm to energy saving design of typical buildings,experimental results show SMOPSO/D can obtain a Pareto optimal solution sets with good convergence and distribution.The thesis has 31 figures,36 tables,and 121 references.