Research Of Some Problems In Artificial Metabolic Algorithm

Along with deeply development in artificial intelligence science, more and more artificial intelligent algorithms are proposed. The historic progress of traditional artificial intelligent algorithms are reviewed and summarized in the paper. The concept and development trend of metabolic engineering in physiology are discussed deeply. Catalysis features, activity and reaction mechanism about enzymes are discussed in detail. The main kinds of enzyme kinetics are overviewed. This study builds artificial metabolic algorithm in the background of biological metabolism function by discussing metabolism rules in physiological science. Similar to traditional genetic algorithm, artificial immune algorithm and artificial neural network algorithm, artificial metabolic algorithm is also presented by simulating biological characteristics. Based on analogy metabolic function, the process of optimization searching for objective function can be regarded as the stage of matching Along with deeply development in artificial intelligence science, more and more artificial intelligent algorithms are proposed. The historic progress of traditional artificial intelligent algorithms are reviewed and summarized in the paper. The concept and development trend of metabolic engineering in physiology are discussed deeply. Catalysis features, activity and reaction mechanism about enzymes are discussed in detail. The main kinds of enzyme kinetics are overviewed. This study builds artificial metabolic algorithm in the background of biological metabolism function by discussing metabolism rules in physiological science. Similar to traditional genetic algorithm, artificial immune algorithm and artificial neural network algorithm, artificial metabolic algorithm is also presented by simulating biological characteristics. Based on analogy metabolic function, the process of optimization searching for objective function can be regarded as the stage of matching operating among metabolites. The matching degree can be adjusted by metabolic operators continuously. Following with the matching degree reaching ideal value, the optimization operation for objective function is also completed. Artificial metabolic algorithm enriches and develops the kind of artificial intelligent algorithms. It provides a new research method and a new research field for intelligent science.The main work in the paper includes:First, the main performance indexes of artificial metabolic algorithm are built. Based on enzyme catalytic mechanism, metabolic competition operator is designed by researching competition match theory between substrates and enzymes. The capability to search global value can be improved by changing competition threshold value. Metabolic apoptosis operator is designed by researching the ability of enzyme to find fitting substrate because of the low matching degree. Searching space for global value can be enlarged by changing apoptosis threshold value and the probability of falling into local extreme value in optimization process can be decreased. Metabolic equilibrium operator is designed by researching the balance theory among metabolites in metabolic reaction process. The quality of producing new individual and keeping suitable convergence speed can be kept by equilibrium operator.Second, artificial metabolic algorithm is designed completely based on analyzing and researching the integrated functions of metabolic competition operator, metabolic apoptosis operator, and metabolic equilibrium operator. Code rules are proposed also. The code patterns in artificial metabolic algorithm are discussed. Artificial enzyme identification theorem and algorithm convergence are proved. The performance parameters in finding optimal value between artificial metabolic algorithm and traditional genetic algorithm are compared according to simulation results in testing functions in detail. The advantages of artificial metabolic algorithm can be displayed. The functions of metabolic parameters in the process of finding optimal value are analyzed in detail. Good reference significance of selecting suitable metabolic parameters for solving problem is provided.Third, by researching metabolite crystallization principle, simulated quenching computing mechanism in artificial metabolic reaction process based on dynamic recrystallization theory is proposed. Interactive adjustment between the metabolic product and optimal process can be achieved effectively. By building fuzzy rules in covalent enzyme-substrate state transition probability, the advantages of simulated quenching computing mechanism are described in taking logistics problem as an example.Fourth, by analysis product between substrate and enzyme and research working principle of apoptosis operator, the method for fault diagnosis through metabolic operating is provided. Fault can be located by the algorithm efficiently. In optimization process, impossible reasons are eliminated continuously by apoptosis operator. The cost for fault diagnosis is decreased and unnecessary search space can be omitted as much as possible. The theory about metabolic traffic flow regulation coefficient and metabolic sensitivity analysis is built. Network system fault diagnosis and self-healing can be realized based on analysising metabolic markers.Fifth, general characters among several algorithms in intelligent science are analyzed. The design mode for computer based on artificial metabolic system is investigated. The application future of artificial metabolic theory in biological science is analyzed. Qualitative discussion for some general problems in intelligence science is given at last.