Optimize The Design Of Software Systems And Its Application

The optimization design is always applied to engineering problems. To optimize an engineering problem, some optimization techniques are adopted to search an optimal solution. The concept of developing an optimization algorithms library (OAL) is conceived. Such a library should contain many optimization algorithms to solve different problems, to be applied on different situations or at specific stages. One can choose an appropriate optimization algorithm comprehended in this library to optimize a problem. To satisfy these requirements, this paper develops an optimization design software system, which includes an OAL. The software system can be applied to the optimization design of aerodynamic shape for reentry vehicles (ASRV).The software system is developed on Window2000 platform. And it consists of two parts. One part is the OAL, the other part is graphic user interface (GUI). The former includes three types of design optimization methods. They are 1) numerical optimization algorithms, 2)logic-based methods, 3)knowledge-based techniques. The development of numerical optimization algorithms and logic-based methods adopts object-oriented strategy. So, these methods provide a unified interface to the client programs. And it is easily to extend the OAL. Knowledge-based optimization method focuses on the directed heuristic search (DHS). DHS employs the knowledge of specific domains, acquired experiences and the known properties of the optimization problem to direct search process. From this, DHS can find a solution to the optimization problem quickly. And for the consideration of the fact that no one single optimization method is good for all problems, hybrid optimization plan (HOP) is disucssed. By integrating more than one single optimization methods together, HOP can maximize the benefits of each method, and compensate for each other's weakness as well. The implementation of HOP is to form a sequence of an arbitrary combination ot'pre-stored optimization techniques, switching from one to another; the sequence can be cycled repeatedly within a single design. And the optimization software system has a convenient GUI. GUI provides an optimization wizard, which directs to use the optimization system. The wizard also allows the user to change the run-time arguments of the optimization method he selects. So the user can control the optimization process.The paper takes the aerodynamic shape optimization design of a reentry vehicle with cruciform flaps as an example of applying the optimization software system toengineering problems. The optimization results verify the effectiveness, the validity and the practicability of the algorithms presented in the paper.