【正文】 nalysis has proven to be the most frequently used method of identifying and solving the problems associated with these plicated designs. Because most of the design tasks in engineering are quantifiable, puter have revolutionized the highly iterative design process, particularly the procedures for quickly finding alternative designs. But even now, many engineers still follow a manual trialanderror approach. Such an approach makes designingeven for seemingly simple tasks more difficult because it usually tasks longer, required extensive humanmachine interaction, and tends to be biased by the design group’s experience. Design optimization, which is based on a rational mathematical approach to modifying designs too plex for engineer to modify, automates the design cycle. If automated optimization can be done on a desktop platform, it can save a lot of time and money. The goal of optimization is to minimize or maximize an objective, such as weight or fundamental frequency that is subject to constraints on response and design parameters. The size and /or shape of the design determine the optimization approach. Looking at optimizations as part of the design process makes it easier to understand .The first step includes preprocessing, analysis, and post processing, just as in customary finite element analysis(FEA) and puteraided design (CAD) program applications (the difference in CAD lies in building the problem’s geometry in terms of the design parameters).In the second step, the optimization objective and response constraints are defined. And in the last step, the repetitive task of design adjustment is automated. Optimization programs should allow engineers to monitor the progress of the design, stop it if necessary, change the design conditions, and restart. The power of an optimization program depends on the available preprocessing and analysis capabilities. Applications for 2D and 3D need both automatic and parametric meshing capabilities. Error estimate and adaptive control must be included because the problem’s geometry and mesh might change during the optimization loops. Revising, remeshing, and reevaluating models to achieve specific design goals start with preliminary design data input. Next es the specification of acceptable tolerances and posed constraints to achieve an optimum, or at least improvement , solution. To optimize products ranging from simple skeletal structures to plicated threedimensional solid models, designers need access to a wide variety of design objectives and behavior constraints. Additional capabilities will also be needed for easy definition and use of the following : weight, volumes, displacements, stresses, strains, frequencies, buckling safety factors, temperature, temperature gradients, and heat fluxes as constraints and objective functions. Moreover, engineers should be able to bine constraints from different types of analyses in multidisciplinary optimization. For example, designers can perform thermal analysis and transfer temperatures as thermal loads for stress analysis, put constraints over maximum temperature, maximum stress, and deflection, and then specify a ranger for the desired fundamental frequency. The objective function can represent the whole model or only parts of it. Even more important ,it should reflect the importance of the different portions of the model by specifying weight or cost factors. Computeraided design (CAD) and puteraided manufacturing (CAE) use digital puter, with their high speed and accuracy, as integrating force throughout the entire process from engineering design to product manufacture. CAD/CAM evolves from and brings together such technologies as numerical control and puter technology. Tied together by the mon usage of digital data, and fed by a continuous stream of electronics developments, numerical control and the digital puter have reached the maturity to permit a level of integration that brings us to the threshold of the longenvisioned automatic factory. The various puterized functions under CAD/CAM fall into three general areas:1. Design/drafting, or puteraide engineering. 2. Planning/scheduling, or management information systems. 3. Fabrication, or manufacturing automation.。A conceptualized or idealized CAD/CAM system is shown in Fig. This shows how the user interacts with the puter via a graphics terminal, designing and controlling the manufacturing process from start to finish with information stored in a shared data base. With the advent of interactive graphics, the problem of the user entering data and instructions into the puter with stacks of coded punched cards or reels of taps went away. No longer was the user required to be experienced in puter programming and operation to use the machine. With interactive graphics, the users municates with the puter in displayscreen picture. No knowledge of puters is required to operate these systems, and the munication is in real time, which means that the puter’s response to the user’s instructions is almost instantaneous. In CAD/CAM, the requirement is for the solution of threedimensional mechanical design and manufacturing problems. Interacting with the puter via keyboard and light pen or other pencil like devices, the designer specifics points and lines on the display screen to quickly construct a displayscreen drawing or model. This is in reality the representation of the diagram stored in the puter data base.With a stroke of the pen or by pushing a button, the designer can move, magnify, rotate, flip, copy, or otherwise manipulate the entire design or any part of it. For example, by pushing a button the designer may issue a flip mand may be used to create models for linear parts. A cross section may be defined and then translated linearly to create a surfaced model. In a similar manner, circular parts can be easily modeled with a rotate mand in which a cross section is rotated about a centr