基于灰关联的PSO-BP神经网络的高层住宅造价估算

针对高层住宅工程造价管理的难点及传统造价估算方法存在的不足，采用灰关联分析与粒子群优化的 BP 神经网络相结合的方法，以高层住宅工程特征指标为网络的输入向量，达到快速、准确地估算高层住宅工程造价的目标。借助文献回顾法与灰关联分析法系统地确定工程特征指标体系并作为神经网络的输入向量；引入 PSO 算法优化 BP 网络的权值及阈值，解决网络收敛速度慢、易陷入局部极小值等缺点。并通过实例验证构建的模型，提高了前期决策阶段造价估算的精确度，实现了快速估算

High-rise Residential Cost Estimation Based on Grey CorrelationAnalysis and PSO-BP Neural Network

In view of the difficulty in cost management of high-rise residential buildings and the deficiency of the traditional costestimation method，this paper proposes a method combining gray correlation analysis and BP neural network of particle swarmoptimization. The characteristic indexes of high-rise residential buildings are used as input vector of the neural network in order toestimate the cost of high-rise residential buildings rapidly and accurately. Firstly，the literature review and the grey relational analysismethod are used to systematically establish the characteristic index system, which is subsequently used as the input vector of theneural network. Secondly，the PSO algorithm optimized BP network weights and thresholds are introduced to solve the disadvantagesof slow convergence and local minima in network. Finally，the model is verified by an example to improve the accuracy of the costestimation in the early stage of decision-making to achieve a quick estimation.