Differential Evolutionary Algorithm And Its Application

With the rapid development of science and technology, a lot of information and a variety of optimization problems are generated in research and engineering areas. Analysis and optimization for these problems has become a common requirement in various fields, such as the train energy-efficient operation, portfolio optimization, optimization of production scheduling and so on. Differential evolution algorithm as an artificial intelligence optimization algorithm is widely used. Because of its ability to solve the discontinuous and non-differentiable function, differential evolution algorithm is gradually attracted the interest of many scholars in recent years. The energy saving of train and stock portfolio optimization problems have been studied in this paper.Energy-efficient operation techniques have been paid more and more attention in subway systems for reducing the cost of operation companies and emissions to the environment. Draw the conclusion that the optimal driving strategy is consist of maximum acceleration, cruising, coasting, and maximum braking. We study the driving strategies during interstations based on the distance and total trip time have been given. Evolutionary algorithms are used as tools to obtain the speed profile with minimum energy consumption. The variable gradients, variable speed limits, maximum acceleration and deceleration, and the maximum traction and maximum braking force varying with train speed were taken into consideration to meet the practical situation. In addition, two cases are studied to validate the effectiveness of proposed method. The experimental results show that the method proposed in this paper can effectively solve the problem of energy-saving of the train, which has a certain practical significance.In order to enable investors to obtain the scientific investment decision, the searches on stock portfolio optimization have been studied. For the constraint that the summation of every stock's weight is 1 in Markowitz's(Markowitz) mean variance model, a new process method was proposed. The weights of stocks are random initialization in the range of variable range, keep the weights value until the sum is 1 in descending order, the remaining stock's weights are reset to 0. Solving the problem that there have too many small positions for investors hardly to handle well in use the traditional method. In addition to, a assets pre-selection process based on non-dominated sorting and related financial theory is proposed to solve the problem of the large-scale portfolio optimization. The new methods for Markowitz theory combined with Multi-Objective Differential Algorithm(MODE) were used to optimize and analyze the investment portfolio of 200 and 1000 stocks in Chinese financial market. The experimental results show that the constraint handling methods can solve the stock optimization problem feasibility. The simulation time is reduced greatly by using the pre-selection process for all the tested algorithms.