Some Improved Artificial Immune Algorithms And Their Applications In The Field Of Atmospheric Environment

Some theoretical study and applications study on immune clonal selection algorithm (ICSA) based on artificial immune system are made in this paper. Moreover the proposed and improved immune clonal selection algorithms are applied to the fields of assessment and forecasting atmospheric quality, and some satisfactory results are also obtained. The major contents and methods could be summarized as follows:(1) The immune clonal selection algorithm introduced into vaccination strategy (ICSA-VS) is proposed in the paper. In the course of the vaccine extraction, selection and vaccination, the roulette selection algorithm is introduced into the ICSA-VS algorithm. Meanwhile some other methods are also proposed such as binary digit gene bit selection and vaccination strategy etc.The float coding is used to realize the ICSA-VS algorithm in the paper. According to the complex degree of problems and the real needs confirm the clone scale of antibodies. The vaccine extraction regards the all antibodies in the choiceness antibodies aggregation as the candidate vaccine population. According to the rate of individual vaccine affinity occupies the sum of the candidate vaccine affinity, so the selected probability for the individual candidate vaccine can be obtained. Then the candidate vaccines are selected by roulette approach. The selected vaccines and the cloned antibodies crossover according to the gene bit obtained by binary digit gene bit selecting and form the new antibodies. The proposed ICSA-VS algorithm has many advantages such as randomness, self-adaptive and diversity and so on, which could improve the vaccination probability of the choiceness antibodies and vaccines, ensure them to inherit and continue in the filial generations and realize the immune self-adjusting function.(2) The immune clonal selection algorithm introduced into local Gaussian mutation operator (ICSA-LGMO) is proposed in the paper. The Gauss distribution has many excellent characteristics such as concentricity, symmetry, uniform variability and so on. Some improvements are made based on ICSA-LGMO algorithm in the paper. The ICSA-LGMO algorithm is introduced into local Gaussian mutation operator to guidance antibody genes variation. By the step of local Gauss mutation self-adaptive adjusting constantly, it could realize that the antibodies and genes wobble in the local area, so that the new antibodies and genes could be searched to replace the initial antibodies and genes, thus the new antibodies produced. The proposed ICSA-LGMO algorithm in the paper could improve the local solutions precision effectively and overcome the local search weak lacking of traditional immune clonal selection algorithm.(3) In view of the advantages of ICSA-VS algorithm and ICSA-LGMO algorithm, the two improved algorithms are combined each other in the paper. Thus the new immune clonal selection algorithm introduced into vaccination strategy and local Gaussian mutation operator (ICSA-VSLGMO) is proposed. In addition, the course of the two algorithms reinforce, promote and improved each other is analyzed in the paper. The local search performance is more exquisite and the precision of solutions is greatly improved in the ICSA-VSLGMO algorithm. The improvement of solution precision can not realize only by adding run time and improving iteration number, but it is the supplement, accelerating and improvement each other for the advantage of two algorithms.In addition, some other strategies such as expanding research space, tabu algorithm and insertion sort etc. are also applied in the improved immune clonal selection algorithms, which it could ensure to obtain the better new antibodies and genes in the whole search area, avoid evidently getting into the local peak value and improve the efficiency of the improved algorithms.(4) Base on ICSA-VSLGMO algorithm, according to the objective function of the assessment atmospheric quality, the ICSA-VSLGMO algorithm is used to optimize the parameters in the formula of pollution harm rate for atmosphere quality assessment. Thus a novel assessment model and a new assessment method based on immune clonal selection algorithm are proposed. The obtained experimental results by simulating experiments are compared and analyzed, they show that the searching capability of the proposed ICSA-VS, ICSA-LGMO and ICSA-VSLGMO algorithms are greatly improved compared with the traditional immune clonal selection algorithm, and the precision of solutions is also improved. The proposed ICSA-VSLGMO algorithm could improve effectively the convergence speed meanwhile maintaining population diversity. The assessment method has the characters of pellucid principle and physical explication. Moreover, the accurate assessment results are also the important advantage of the method. It could provide a new effective approach for artificial immune theory and technology applied to atmospheric environment field. So it has better practicability and application foreground.(5) A new dynamic threshold strategy that suits to immune clonal selection algorithm is proposed in the paper. In the method, the similarity among antibodies confirms the initial threshold. The decreasing function of threshold is proposed and it is applied to constrain the decay amplitude, which could avoid producing the similar antibodies. It could overcome evidently the consanguineous reproduction generated by the immune clonal selection algorithm optimizing the muti-parameters and premature convergence.(6) Owing to the ICSA-VSLGMO algorithm introduced into dynamic threshold strategy has better advantage on optimizing multi-parameters, therefore a novel approach that the improved immune clonal selection algorithm, named as ICSA-VSLGMO algorithm introduced into dynamic threshold strategy, is used to optimize the dynamic recursion neural network is proposed in the paper. The concrete structure of the recursion neural network, the connect weight and the initial values of the contact units etc. are done by evolving training and learning automatically. Thus it could realize to construct and design for dynamic recursion neural networks. In addition, we regard the dynamic recursion Elman neural network as a case to analyze in the paper. It could provide a new effective approach for immune clonal selection algorithm optimizing dynamic recursion neural networks.(7) The Elman neural network optimized by the improved immune clonal selection algorithm, the double-feedback Elman neural network introduced into direction information and the basic Elman neural network are all applied to forecast atmospheric quality. The obtained experimental results by forecasting SO2,NO2 and PM10 are compared and analyzed show that the improved Elman neural network optimized by the improved immune clonal selection algorithm has pretty well fitting and forecasting capability. It is applied to forecast the atmospheric quality of someone city in Jilin Province and some satisfactory results are also obtained, which has better application value and application foreground.