Automatic Design Of Cell-Like Polynomial Membrane Systems

As a new branch of natural computing, membrane computing means to abstract computing models from the structure and function of living cells, as well as from the way the cells are organized in tissues or organs. Also known as P system, membrane computing model earns wide concern with its advantages such as distribution and maximum concurrent computation.Some membrane computing models have been proposed based on theoretical research with the help of mathematic and form language. But most of these models were proposed based on designers’experience and demonstration, it took too much time and energy. Therefore, combining evolutionary algorithm and the design of membrane computing model to achieve automatic design has become a research hotspot. Based on the recent results, there have been some automatic designed P systems who are able to complete some calculation tasks, such as42, n. However these models can only calculate simple polynomials, they haven’t formed a unified computing framework which will influent the development of automatic design. In this case, this paper took cell-like membrane system as the research object, by using genetic algorithm and simulation software P-Lingua completed the automatic design of the polynomial P systems, the main works and achievements are as follows:1. Completed a software wirten by C#which was used to calculate the types of membrane systems. Introduced the tread of design, development process, function, usage and used an example to verify the effectiveness of the software.2. Proposed a method for the polynomial membrane system. Introduced the permutation encoding, the method to judge a membrane system is terminated or not, the improved evaluation method of a function with penalty, the selection of operators of genetic algorithm. The experimental results showed this presented design method was effective.3. Proposed the automatic design method of membrane systems which have variable number of rules, introduced the thread of method, the pocess of how to deal with the sets of the evolution rules, and applied this method to design membrane system for n2and the polynomial. Experimental results showed that the number of evolution rules can be variable within the scope of the biggest rule’s number. Compared with the previous methods, this method can design a fitable membrane system which has less rules, and it will lay the foundation for finding the minimum membrane system.