Investigation Of Evolutionary Algorithms And Its Applications For Information Engineering

Computation Intelligence (CI) is such a theory and method to obtain, process and apply useful information based on simulating human's intelligent mechanism or living's evolution processing or human (or other species)'s intelligent behavior. This paper has presented some theoretical and application studies in the Evolutionary Algorithms (EAs). The main research result can be summarized as follows:In chapter 1, we give a brief discussion about the origin and development of CI. Its characteristics, applications and research status were generalized. This is followed by a guide to this structure of paper.A novel evolutionary algorithm, Uniform Design Genetic Expression Programming (UGEP), is proposed in the character 2 based on the conception of Uniform Design. Compared with standard GEP, UGEP employs a new strategy to create initial population owning to mixed-level uniform array and uses a novel and robust crossover based on uniform design to produce the next generation. Theoretic analysis proves that UGEP converges to the global optimum and its convergence speed is faster than standard GEP.Combined with the concept of the virus evolution principles, a novel algorithm, virus-Gene Expression Programming (VGEP), is proposed in section 3 to deal with both funtion fitting and time-series optimization problems. It can improve searching efficiency and decrease the probability of premature phenomena by constructing a new update and infect mechanism for virus. And also the results showed the effectiveness of new algorithm.The problem of Minimum Weight Triangulation (MWT) is one of the most important issues in computer vision. An adaptive Immune Genetic Algorithm (AIGA) to solve the problem is proposed in character 4. Simulation results show that the proposed algorithm performs better in terms of quality of MWT, especially for the large scale of point cluster, and has good practical value.In section 5 a novel method in camera calibration for Binocular Vision with high precision is put forward. Based on the radial distorted analytical model, the approach is presented to partition the camera region and compensate distortion simultaneity. It introduces the SAGEP (self- adaptive Genetic Expression Programming) theory to process pixels from the corresponding region so as to discover valid coordinate model automatically.Finally, we develop a novel GIS to control the moving objects. The system, based on GPS/GIS/GPRS/RS satellite orientations (network edition), has gathered Global Position System (GPS), Geography Information System (GIS) on the basis of communication server. Its main function is to transport the information as the dynamic position of the moving object (longitude, latitude), time and appearance to the main center by the wireless. And then users can search the tracks of object through the system and check the exact position, speed, direction and state of the moving object which the users are interest in. What is more, owning to that system, users can supervise and control the track and appearances of the moving objects conveniently, benefiting economic and society obviously. The experiment has showed that the project can solve many problems in the same kind systems, which has got a very wide application.