Building energy consumption prediction method based on artificial bee colony algorithm and neural network

Abstract

The invention provides a building energy consumption prediction method based on an artificial bee colony algorithm and a neural network. The method comprises the steps that firstly, the artificial bee colony algorithm is utilized for conducting weight value optimization on the neural network; secondly, the optimized neural network is utilized for predicting building energy consumption. The artificial bee colony algorithm is an optimizing algorithm simulating a bee colony and has the advantages that control parameters are fewer, implementation is easy, and calculation is convenient; compared with a particle swarm algorithm, a genetic algorithm and other intelligent computing methods, the artificial bee colony algorithm has the prominent advantages that in each iterative process, global search and local search are both performed, the probability of finding an optimal solution is greatly increased, local optimum is avoided to a great extent, and global convergence is enhanced. Thus, when the artificial bee colony algorithm is adopted to optimize the initial weight value of the neutral network, the accuracy of the neutral network predicting the building energy consumption is improved, and meanwhile the defects existing in weight value optimization of the neutral network at present can be overcome obviously.

Description

technical field [0001] The invention relates to a method for predicting building energy consumption, which belongs to the field of building energy consumption prediction, in particular to a method for predicting building energy consumption based on artificial bee colonies and neural networks. Background technique [0002] With the rapid development of the economy, the scale of my country's buildings is getting larger and larger, and the problem of high energy consumption is becoming more and more serious. It is of great significance to the energy-saving planning of buildings to do a good job in energy-saving management and energy-consumption monitoring. Building energy conservation is the focus of today's urban construction, and the prediction, analysis and evaluation of building energy consumption is the premise and basis for realizing building energy conservation. Neural network is a typical intelligent computing method. Because the algorithm has simplicity, plasticity, hi...

AI Technical Summary

Problems solved by technology

For example, in the article "Study on the Prediction Method of Support Vector Machine Based on Ant Colony Optimization Neural Network", the author uses the ant colony algorithm to train the weight and threshold of the neural network, thus overcoming the shortcomings of the low prediction accuracy of the neural network, but the ant colony algorithm After searching to a certain extent, the solutions found by all individuals are exactly the same, and the solution space cannot be further searched, which is not conducive to finding better solutions.

In the article "Application of 96-point Load Modeling of Combined Particle Swarm Neural Network", the author aimed at the long training time of the original BP neural network and the lack of easy to fall into local minimum values, and integrated the particle swarm algorithm into the neural network, using Particle swarm has the advantages of fast training speed and high precision. Training the weights and thresholds of the neural network can effectively improve the generalization ability and learning ability of the neural network, and improve the prediction accuracy of the neural network. However, the overall situation of the particle swarm algorithm itself is Convergence needs to be further improved

In the article "Application of Genetic Algorithm in Neural Network Weight Optimization", the author uses genetic algorithm to optimize the weight of neural network, which can effectively overcome the shortage of falling into local minimum and improve the global search ability, but the training time is longer , which is not conducive to the application of neural networks

[0004] In the above technologies, although the weight optimization of the neural network with ant colony algorithm, particle swarm algorithm and genetic algorithm can overcome the problem that the neural network is easy to fall into the local minimum to a certain extent, it can effectively improve the generality of the neural network. ability and learning ability to improve the prediction accuracy of the neural network, but there are also obvious deficiencies

During the optimization process of the ant colony algorithm, when the search reaches a certain level, the solutions found by all individuals are exactly the same, and the solution space cannot be further searched, which is not conducive to finding a better solution.

In the optimization process of the particle swarm optimization algorithm, its own global convergence is not very good, and needs to be further improved

However, the genetic algorithm takes too long to optimize the neural network, which is not conducive to practical application.

Images