### Turbojet Engine Performance Modelling using Multi-Objective Optimization Algorithms

Abstract

Many real world problems are multi-objective in nature. There are many
methods that solve this kind of optimization problem. They can be classi-
fied into two classes: genetic algorithms and classical methods. Instead of
one solution, both methods find a whole set of solutions. Genetic algorithms
seem to attract many researchers due to their robustness and also to their
degree of generality, while others prefer methods based on classical optimization
algorithms. The choice of one approach over another depends much on
the nature of the problem. If we have enough information about the optimization
problem, it is better to use methods based on classical optimization
algorithms. If not, it is better to use multi-objective genetic algorithms or
evolution algorithms. In this thesis we consider a design optimization problem.
The task is to design a jet engine by specifying certain parameters such
as pressure ratio in the turbine. We run a flight simulation in Matlab, which
allows us to calculate the fuel consumption as well as the engine weight for
different input parameters. Here the input parameters are decision variables
while the objective functions are the fuel consumption and the engine weight.
We applied two multi-objective evolution algorithms and one method based
on classical optimization algorithms.