The Artificial Life and Adaptive Robotics
Laboratory - ALAR
Downloadable Archives
1. Evolutionary Multi-objective
Optimization
- EMO in Dynamic Environment
+
EMO: Binary version of NSGA2
+
Method: EMO is used to solve the single objective dynamic problems. The
artificial objectives are created to
transform single-obj problems to multi-obj ones.
+
Test problem: Moving Peaks Benchmark.
+
Reference paper:
Bui L.T., Juergen B. and Abbass
H.A. (2005) Multi-objective optimization for dynamic environments. The IEEE
Congress on Evolutionary Computation (CEC),
+ Download: C++ Source Code
(see readme.txt for
details)
- EMO in Noisy Environment
+ EMO: Binary versions of NSGA2
and SPEA2
+ Test problems: ZDT1, ZDT2, ZDT3,
ZDT4, ZDT5, ZDT6
+ Experiments: Noise is
additive to the real fitness values with difference levels.
+ Reference Paper:
Bui L.T., Essam D., Abbass H.A.,
Green D. (2004) Performance analysis of evolutionary multi-objective optimization
algorithms in noisy environments, Asia Pacific Symposium on Intelligent and
Evolutionary Systems, pp. 29-39
+ Download: C++ Source
Code (see readme.txt for details)
2.
Data Mining
- Distributed Data Mining using XCS
+
Continuous versions of XCS
+
Test problem: real multiplexer, checkerboard
+
Experiments: A distributed XCS system for data mining. The system consists of a
number of clients and a server. A client is placed at each distributed site and
is responsible for gathering local knowledge and transferring information back
to the server. The server combines the information from the clients to form a
descriptive model for the global environment.
+
Reference papers:
H.H. Dam, H.A. Abbass, C.J. Lokan, The
performance of the DXCS System on Continuous-Valued Inputs in Stationary and
Dynamic Environments. CEC 2005, (IEEE Congress on Evolutionary Computation).
H.H. Dam, H.A. Abbass, C.J. Lokan, DXCS:
An XCS System for Distributed Data Mining. GECCO 2005, (Genetic and
Evolutionary Computation Conference).
+
Download: C++ Source Code (see readme.txt for details)
- XCS for Dynamic Environments
+
Continuous versions of XCS
+
Test problem: real multiplexer
+
Experiments: XCS is explored in dynamic environments with different magnitudes
of change to the underlying concepts.
+Reference
papers:
H.H. Dam, H.A. Abbass, C.J. Lokan,
Evolutionary Online Data Mining – an Investigation in a Dynamic
Environment. 2005, accepted for a book chapter in Springer Series on Studies in
Computational Intelligence
H.H. Dam, H.A. Abbass, C.J. Lokan, Be
Real! XCS with Continuous-Valued Inputs. IWLCS 2005, (International Workshop on
Learning Classifier Systems).
+ Download: C++ Source Code (see readme.txt for details)
- 1999 KDDCUP FTP-Only
Datasets
o
KDD FTP Training set (Normalized)
o
KDD FTP Test set (Normalized)
3.
Co-evolution
- Co-evolution of Genotype-phenotype mappings (GPM) to solve epistatic
optimization problems
+
Method: Using cooperative co-evolution mechanism to adapt GPM in order to
facilitate the search of optima in epistatic optimization problems. The chief
objective is to learn the problem structure while searching for the optima.
+
Test algorithm: Fast Evolutionary programming
+
Test problem: Rotated Rastrigin with 10 variables.
+
Reference Paper:
Bui L.T., Abbass H.A., and Essam D.
(2005) Cooperative Coevolution of Genotype-Phenotype Mappings to Solve
Epistatic Optimization Problems, in Advances in Artificial Life, Abbass H.A.,
Bossamier T., and Wiles J. (Eds), World Scientific Publisher, pp. 29-42.
+
C++ Source
Code (see
readme.txt for details)
4. Genetic Programming
- AntTAG:
Ant Colony Optimisation (ACO) inspired method to synthesise programs
+ Method: ACO inspired search method
is used to explore grammar constrained search space to synthesise tree structures
resembling the representations in Genetic Programming.
+ Test problem: Symbolic regression
problem X^4+X^3+X^2+X and X^5+X^4+X^3+X^2+X
+ Reference Paper:
Abbass
H.A., Hoai N.X., and
+ Download
: C++ Source Code
5. Multi-agent
systems
-
Military Operations simulation: WISDOM (beta release)
Last
update : 25/11/2005