Information Technology Reference
In-Depth Information
omni-aiNet: An Immune-Inspired Approach for
Omni Optimization
Guilherme P. Coelho and Fernando J. Von Zuben
Laboratory of Bioinformatics and Bioinspired Computing - LBiC
Department of Computer Engineering and Industrial Automation - DCA
School of Electrical and Computer Engineering - FEEC
University of Campinas - UNICAMP
PO Box 6101, 13083-970 Campinas, SP, Brazil
{
gcoelho, vonzuben
}
@dca.fee.unicamp.br
Abstract.
This work presents omni-aiNet, an immune-inspired algo-
rithm developed to solve single and multi-objective optimization prob-
lems, either with single and multi-global solutions. The search engine
is capable of automatically adapting the exploration of the search space
according to the intrinsic demand of the optimization problem. This pro-
posal unites the concepts of omni-optimization, already proposed in the
literature, with distinctive procedures associated with immune-inspired
concepts. Due to the immune inspiration, the omni-aiNet presents a pop-
ulation capable of adjusting its size during the execution of the algorithm,
according to a predefined suppression threshold, and a new grid mecha-
nism to control the spread of solutions in the objective space. The omni-
aiNet was applied to several optimization problems and the obtained
results are presented and analyzed.
1
Introduction
During the last decades, the optimization field has been benefited from the con-
tinued sprouting of e
cient optimization algorithms. These algorithms have been
applied to an expressive number of different real-world problems, leading to very
encouraging results. However, optimization problems appear in different types
and forms: some may have a single objective (known as single-objective op-
timization problems); some may have multiple conflicting objectives (known as
multi-objective optimization problems); some problems may have only one global
optimum, requiring the task of finding this optimum; and other problems may
contain more than one global optimum in the search space, thereby requiring the
task of simultaneously finding multiple global optimal solutions. This variability
in features and objectives guided to the proposition of algorithms specialized in
each kind of problem, what forced users to know different algorithms in order to
solve different kinds of optimization problems.
A straight attempt to revert this tendency was made by Deb and Tiwari
[8]. In their work, they propose and evaluate a single evolutionary optimization
algorithm for solving different kinds of function optimization problems: single or
