The Research Of Coating Design Method Based On Intelligent Algorithms

With the development of optical technology, optical coating design is required for an increasingly higher requirement in optical devices. The designing of optical film plays an important role in performance of optical devices. The merit function in coating design is used to review the quality of the optical properties of optical devices. The merit function in coating design is a multivariate function composed of the film structure parameters. Coating design is being completed through approaching the optimal solution of such merit function with numerical method. The mentioned merit function mentioned here considers variables including film structure, layers, thickness and index of refraction so it is highly challenged to be analyzed. Moreover, the complexity of merit function is proportional to the layers. It is very difficult work for traditional optimization techniques such as Monte Carlo method, simplex method etc. to find the global minimum.Intelligent optimization algorithms become the centre of attention in the area of coating design. Intelligent optimization algorithms imitating nature law are a branches of bionics. For example, ant colony algorithm(ACA), genetic algorithm(GA), particle swarm optimization(PSO), differential evolution(DE) are the popular technologies in dealing with global optimization problems.The main points we will concern in this thesis is applying PSO and DE in coating design. Using MATLAB language to compile algorithm program and apply them to coating design such as anti-reflection film, high-reflection film, beam-splitting film, filter, far infrared antireflection coating. In the examples, the sum of squared error of ideal reflectivity and actual reflectivity was used as the merit function of multilayer coating performance. Detailed comparison of performance are processed for every intelligent algorithm in coating design problem, so as to find the most suitable one. The results indicate that the PSO and DE are effective techniques for the design of multilayer coatings. By means of PSO or DE, a better multilayer coating structure than GA can be obtained under the same design condition.