STUDY TO INCREASE THE PERFORMANCE OF THE ADAPTIVE VERSION OF THE GEO ALGORITHM AND ITS APPLICATION IN THE CONCEPTUAL PROJECT OF SPACE SYSTEMS

National Institute for Space Research - INPE

Engineering and Space Technology - CSE/ETE

This repository contains the algorithms developed during a master's thesis. The main focus of the work was to improve the evolutionary algorithm A-GEO, which was presented in the paper A new Adaptive Evolutionary Algorithm for Design Optimization (Barroca, 2019).

• Objectives   • Algorithms  

 

Objectives

• Implement the parameter control mechanism of A-GEO in GEOvar, a variant of GEO, verifying this implementation using a set of test functions;
• Change the encoding of the A-GEO design variables from binary to real, no longer requiring the definition of the number of bits encoding each variable;
• Study how the design variables are perturbed in the real encoding, as well as how to perform several mutations on each variable in a single iteration;
• Investigate the control mechanism for the parameter τ proposed for A-GEO, which may indicate other ways to control the parameter during search;
• Explore ways to make A-GEOreal fully adaptive, without the need for parameter tuning;
• Apply the improved algorithm in the search for automating the creation of solutions in a conceptual design of space systems;

 

Algorithms

Binary Encoding:

- GEO
- GEOvar
- A-GEO
- A-GEOvar
- A-GEO2var_5

Real Encoding:

- GEOreal1 (GEOreal1_M, GEOreal1_P, GEOreal1_A)
- GEOreal2 (GEOreal2_M_VO, GEOreal2_P_VO, GEOreal2_A_VO, GEOreal2_M_DS, GEOreal2_P_DS, GEOreal2_A_DS)
- GEOreal2_P_DS_UNI
- A-GEO2real1_AA
- A-GEO2real2_AA0
- A-GEO2real2_AA1
- A-GEO2real2_AA2
- A-GEO2real2_AA3