An Artificial Bee Colony Algorithm Based On Membrane Computing And Its Application

The artificial bee colony(ABC)algorithm is an optimization algorithm for simulating biological swarm intelligence,which has strong adaptability to solve the optimization problem of continuous functions.At present,it has been widely used in machine learning,combination optimization,data mining and many other fields.However,the ABC algorithm still have some shortcomings such as slow convergence and easy to fall into local optimum.Membrane computing is a branch of biocomputing,and its model is called membrane system or P system.The P system has some advantages such as distributed and parallel computation,non-deterministic and expandable features,and easy implementation.In this paper,in order to overcome the defects of ABC algorithms,evolution and comunication mechanism of P system is used to propose a P system-based artificial bee colony algorithm,called PABC algorithm.And then,the PABC algorithm is used to solve data clustering problem and face recognition problem.The main innovations of this paper are as follows:1)An artificial bee colony algorithm based on P system is proposed,called PABC algorithm.A cell-like P system with one level structure is designed to optimize the ABC algorithm,in which ABC rules are used to achieve the evolution of objects in the elementary membrane.The are communication rules between the elementary membranes and the skin membrane are used to exchange and sharing of objects in the system,which can speed up the convergence of the algorithm.2)An k-means algorithm improved by PABC is proposed,called PABC-K algorithm.The algorithm mainly uses PABC to optimize the K-means algorithm and enhance its performance in clustering applications.A cell-like P system with one level structure is designedrealize the PABC-K algorithm.In the elementary membrane,the operation of combining the K-means algorithm with the ABC algorithm is used to evolve the objects.The exchange and sharing of objects in the system are realized by the communication rules.3)The application of PABC algorithm to face recognition is discussed.The feature selection problem in the face recognition system is considered as an optimization problem,and its objectives include maximizing recognition accuracy and minimizing the cardinality of the feature set.The PABC algorithm is used to select the optimal feature subspace for the feature space.