sea surface temperature time sequence prediction method based on PSO double-objective optimization

Abstract

The invention belongs to the technical field of prediction or optimization, and discloses a sea surface temperature time sequence prediction method based on PSO double-objective optimization, which comprises the following steps: optimizing sea surface temperature prediction model parameters by utilizing an improved PSO double-objective optimization algorithm to obtain an optimized parameter combination; and predicting the sea surface temperature by using the obtained optimization parameter combination. According to the method, the PSO algorithm is improved, and the Pareto dominated relationship is adopted to measure the advantages and disadvantages of the solution.The global search capability of the PSO algorithm is increased by utilizing a large inertia weight, bidirectional local searchis carried out on a non-dominated solution set, the local search capability of the PSO algorithm is enhanced, the PSO algorithm has a high optimization capability, and the Pareto frontier of a real solution can be approached. The local search capability of the particles is enhanced, the diversity of the non-dominated solutions is increased, and the number of the non-dominated solutions is maintained by using the crowded distance, so that the distribution uniformity of the non-dominated solutions is improved.

Description

technical field [0001] The present invention belongs to prediction or optimization, such as linear programming, "traveling salesman problem" or "cutting material problem" technical field, in particular to a sea surface temperature time series prediction method based on PSO dual-objective optimization Background technique [0002] At present, the existing technologies commonly used in the industry are as follows: [0003] Sea Surface Temperature Prediction (SSTP) research can be classified into three categories: (1) Statistical prediction methods; (2) Numerical prediction methods; (3) Empirical prediction methods. The SSTP based on the similarity measure predicts SST by analyzing the similarity of sea surface temperature (SST) sequence changes, which belongs to the empirical prediction method. The main idea is to find a sequence similar to the current trend in the historical SST sequence, and then use the historical trend to predict the current trend. [0004] Sea surface t...

AI Technical Summary

Problems solved by technology

[0007] (1) In the prediction of sea surface temperature, the prediction model parameters have a great influence on the prediction performance of the model, and it is impossible to select a better combination of model parameters;

[0008] (2) The PSO algorithm is easy to fall into the local extremum due to the weak global search ability, thus reducing the optimization ability of the algorithm;

[0009] (3) The computational complexity of the NSGA algorithm is relatively high; when the NSGA-Ⅱ algorithm deals with high-dimensional multi-objective problems, the computational complexity of the crowding distance in the high-dimensional space will be relatively high; the SPEA algorithm may make the obtained non-inferior solution incomplete;

[0010] (4) When the traditional particle swarm optimization algorithm generates the initial solution, it mainly uses random initialization. Due to the lack of prior information guidance, it is not conducive for the initial particles to move closer to the optimal solution.

[0011] (5) When traditional PSO performs multi-objective optimization, weights are usually used to weight each objective function. Decision makers often determine the weights through personal experience, which often makes the optimization results deviate from the actual global optimal solution

Falling into a local optimum can easily lead to unsatisfactory optimization results, resulting in poor parameters of the obtained sea surface temperature model and reducing the performance of the sea surface temperature prediction model

Images