Research On Antenna Array Synthesis Of Immune Algorithm

The biologic immune system can recognize and eliminate foreign antigens to protect the stability of the body with the characteristic of highly evolutionary, highly adaptive, highly distributed and self-organizing. The Artificial Immune System (AIS) is an intelligent information processing technology which based on the mechanism of the vertebrate immune system. AIS has been a hot research domain of strong capabilities in pattern recognition, learning and associative memory.Based on biologic immune system, this paper summarizes the bionics mechanism of immune optimization, while clonal selection and immune network theories are discussed detailedly. An immune optimized algorithm based on clonal selection principle is presented with the research of common immune algorithm. Contributions are summarized as follows:Firstly, according to the flow of immune algorithm, transform process of the algorithm is proved as a Markov chain process and the global convergence is proved in theory. Several multi-extremum functions are optimized to testify the convergence of the algorithm, good convergence characteristic is achived and the convergent speed is rapid.Secondly, as a key technology in many areas such as radar, wireless communi-cations and electronic warfare, antenna array is synthesized by immune algorithm. To the optimization, synthesis problem of antenna array is regarded as the optimizing problem of objective function. In the paper, immune algorithm is used to optimize the array space, excitation amplitude and phase for side lobe reduction, null steering and broading. And also, array lattice is successfully optimized for low side lobe in thinned array.Thirdly, Niche double-crossover immune genetic algorithm is presented by adding genetic crossover operator to immune algorithm, while niche idea is combined to the algorithm. Double-crossover operator is applied to enhance the diversity of searching mechanism by using single point and arithmetic crossover operator in odd and even iteration, respectively. Niche idea is applied in the algorithm to avoid the convergency problem in the searching area by dividing which to several niches. Sharing mechanism of niche is used to compose the affinity of the antibodies for avoiding local convergence which restrained the excessed propagation in antibodies of high affinity and advanced the probability of getting to the next generation in antibodies of low affinity. The niche double-crossover immune genetic algorithm is applied in synthesis of antenna array, and muti-nulls, broad nulls are obtained successfully, which obtained the validity of the algorithm.Fourthly, applying the characteristic of ergodic, randomicity and the sensitivity of initial value, chaos mapping is combined with immune algorithm, and Multi-map chaos immune algorithm is presented. There is a farther optimization to the global searching performance of immune algorithm using the ergodicity of chaos sequence and the randomicity with "rules" of chaos disturbance. Two chaos sequences with different rules are applied to control the generation of the antibodies in different periods, which ensure the diversity of antibodies. As a result the searching area is adequacy, earliness phenomenon is conquered. A hybrid mutation method is presented which combining Chaos disturbance and the own mutation method of immune algorithm. As a result the searching efficiency is advanced and the speed of convergence is improved. Well results for antenna array synthesis show the fine global convergence capability of the presented algorithm.