IOSO Optimization Toolkit - Novel Software to Create Better Design
И. В. Егоров, Gennadiy Kretinin, I. A. Leshchenko, Sergey V. Kuptzov · 9th AIAA/ISSMO Symposium on Multidisciplinary Analysis and Optimization · 2002
Optimization has always been around. Designing of new object or system in any case leads to the necessity to provide minimum or maximum value of some indicator characterizing the efficiency of the system. To find optimum solution one can go in two ways: do a series of parameters variations and pick the best solution from the results, or employ numerical optimization algorithms. The latter way is much more preferable for design optimization problems with more than two variables. It allows a designer to find mathematical optimum faster and more accurately. IOSO optimization software package facilitates better design by allowing search for object function optimum in design space taking into account design constraints. All IOSO software packages use response surface technology in the algorithms of search for optimum. Its response surface implements a complimentary set of approximation algorithms. The strategy of response surface building is set up adaptively during search for optimum, depending on features of object being optimized and current state of search. The paper presents main features of IOSO optimization software and examples of some practical problems solution. In particular, it demonstrates the example of solution of real-life optimization problem for parameters of gas turbine engine being developed for a commercial aircraft.