Research On Improved Harris Hawks Optimization And Its Applications

As a new swarm intelligence algorithm,Harris Hawks Optimization is inspired by Harris'eagle swarm hunting behavior.Harris Hawks Optimization has the advantages of simple principle,few parameters,easy implementation,suitable for combinating with other algorithms and outstanding performance on high-dimensional problems.It can be used to solve various optimization problems.There are also problems such as simple setting of key parameters,limited accuracy on low-dimensional problems and excessive reliance on existing individuals.Aiming at the shortcomings of Harris Eagle Algorithm,this thesis studies and improves Harris Eagle Algorithm and applies it to solving optimization problems with practical background.The main work of this thesis is as follows:(1)An modified Harris Hawks Optimization(MHHO)is proposed.In MHHO,the jumping strength of the prey is updated depend on the escape energy of the prey itself,the elite individuals are selected and retained,and the random individuals perform Levy flight,and the neighborhood search operator is introduced to enhance the local search ability.MHHO,HHO,PSO,WOA are compared with same test functions to show the effectiveness;(2)The Harris Hawks Optimization with differential mutation(DHHO)is proposed.In DHHO,the concept of differential evolution is introduced,and Gaussian mutation is added in order to make the algorithm jump out of local optimum.The selection method of the new population combines the fitness value and population diversity into a comprehensive consideration.The optimization experiments proved improvement of the same test function show the effectiveness of DHHO;(3)The improved algorithm is applied to the identification of photovoltaic cells model parameters,by identifying the main parameters of the two equivalent models of photovoltaic cells single and double diodes.Several algorithms were used to optimize the parameters of the data collected under the same light intensity and different temperatures,and the effectiveness of the improved algorithm are shown;(4)Using RBF neural network is used to model the overhead crane.Using the improved algorithms and some other algorithms to optimize the center point and variance of the RBF neural network.Divide the collected data into a training set and a test set,use the training set to train the neural network,and use the test set to detect.The effectiveness of the improved algorithm based RBF-NN has the best performance.