MOEA - User Manual
Hamburg University of Technology
Institute of Control Systems
Andrey Popov
Contents
1 Copyright and Legal Notes 2
2 Files and Options 3
3 Usage Steps 4
Introduction
Many engineering problems are characterized by several, often contradicting objectives,
that have to be satisfied simultaneously. In many cases the standard approach of combin-
ing the objectives to a single one does not provide satisfactory results [1]. Evolutionary
Algorithms are an attractive alternative in such cases [2], since they are a direct paral-
lel search method and as such are able to obtain a Pareto trade-off surface between the
objective function on a single run.
Strengthen Pareto Evolutionary Algorithm 2 (SPEA 2) is a Multi-Objective Evolu-
tionary Algorithm (MOEA) developed by Computer Engineering (TIK), ETH Zurich,
and reported to work well on many benchmark problems [3]. SPEA 2 is originally written
in C, and developed under the PISA framework (Platform and Programming Language
Independent Interface for Search Algorithms).
In order to allow easy use of SPEA 2 under MATLAB and avoid the file-based infor-
mation exchange between the solver and the problem, the Institute of Control Systems,
Hamburg University of Technology, combined SPEA 2 with the DTLZ variation function,
modified the C files to interface with MATLAB and created an additional M file moea.m
for user convenience.
Used abbreviations:
MOEA Multi-Objective Evolutionary Algorithm
SPEA 2 Strengthen Pareto Evolutionary Algorithm 2
PISA Platform and Programming Language Independent Interface for Search
Algorithms
1
