Research And Application Of Two Swarm Intelligence Algorithms

Swarm intelligence algorithm is the stochastic optimization algorithm which is simulated the group behavior of biological in nature. It gives a new way to find the solution of some complex problems such as constraints, nonlinear and the extraction of extreme value. Ant colony optimization algorithm (ACO) and particle swarm optimization algorithm (PSO) are two important methods of swarm intelligence algorithm. In this paper, an in-depth research will be done using these two methods on sequence alignment.On the one hand, the traditional ACO has been applied in sequence alignment, but it is limited to the alignment of similar length sequences and local optimum. Thus developing an accurate gene sequence alignment algorithm remains to be a very difficult computer problem. A new sequence alignment method based on ACO was brought forward in the paper. The method could avoid local optimum; especially remove the wide difference paths' scores by regulating the initial and final position of ants and modifying pheromones in different time. Therefore the method has the ability of aligning sequences of different length and avoiding the local optimal causing by traditional algorithms. The results show that the novel sequence alignment algorithm is efficient and feasible. On the other hand, PSO simulates the birds motion behavior, the character is simple, fast convergence and easy to implement. This paper first introduced the superiority of PSO for solving the problems, and then given a new improved PSO: According to different applications using the corresponding adaptive inertia weight, taking into account the positions of other particles to balance of the current particle's position. The advantages of this improvement is that making particle more conform to the evolution characteristics, helping to speed up the algorithm convergence, improving the search performance of PSO, reducing shortcomings of local optimization and increasing the global convergence capacity. This article was applying PSO on the sequence alignment, given the constraint condition of initial position of particles and moving strategy in the process of particles' flying. Based on the experimental, it proved this method is feasible and effective.This paper compares the improved ACO and PSO with the same data, shows their selves characters. Using sequence alignment to find the fragment is important for the extraction of clean sequence. This paper also applies ACO and PSO to the extraction the fragment of cDNA sequences, proves that the two methods on searching sequence fragment are feasibility; shows ACO and PSO have the great potential and satisfactory optimal performance.