Title :
Visualization of EC landscape to accelerate EC conversion and evaluation of its effect
Author :
Boschetti, Fabio ; Takagi, Hideyuki
Author_Institution :
Div. of Exploration & Min., CSIRO, Nedlands, WA, Australia
Abstract :
We evaluate how visualization of an evolutionary computation (EC) landscape is effective using a geophysical task. This technique allows us to actively participate in EC optimization by viewing the distribution of searching points on 2D space mapped from an n-D EC landscape, and indicating where in the EC is the possible global optimum. We construct a Visualized GA system that includes self-organizing maps for visualization and compare its performance with that of a normal GA using the geophysical simulation task. Sign tests for the comparisons show that the Visualized GA converges significantly faster than the normal EC (p<0.01), which suggests further extensions to enhance user interactivity
Keywords :
data visualisation; evolutionary computation; geophysics computing; interactive systems; search problems; self-organising feature maps; 2D space; EC conversion acceleration; EC landscape visualization; EC optimization; Visualized GA system; evolutionary computation; geophysical simulation task; geophysical task; global optimum; n-D EC landscape; normal EC; normal GA; self-organizing maps; sign tests; user interactivity; Acceleration; Australia; Computational modeling; Evolutionary computation; Humans; IEC standards; Physics computing; Self organizing feature maps; System testing; Visualization;
Conference_Titel :
Evolutionary Computation, 2001. Proceedings of the 2001 Congress on
Conference_Location :
Seoul
Print_ISBN :
0-7803-6657-3
DOI :
10.1109/CEC.2001.934283
