Bartz-Beielstein T. Experimental Research in Evolutionary Computation. The New Experimentalism

Springer, 2006. — 221 p.Rigorously proven upper and lower run-time bounds for simplified evolutionary algorithms on artificial optimization problems on the one hand and endless tables of benchmark results for real-world algorithms on today’s or yesterday’s hardware on the other, is that all one can do to justify their invention, existence, or even spreading use? Thomas Bartz-Beielstein gives thoughtful answers to such questions that have bothered him since he joined the team of researchers at the Chair of Systems Analysis within the Department of Computer Science at the University of Dortmund. He brings together recent results from statistics, epistemology of experimentation, and evolutionary computation.
After a long period in which experimentation has been discredited in evolutionary computation, it is regaining importance. This book far exceeds a discussion of often-met points of criticism of the usual experimental approach like missing standards, different measures, and inaccurate and irreproducible results. Also, fundamental objections against the experimental approach are discussed and cleared up. This work shows ways and means to close the gap between theoretical and experimental approaches in algorithm engineering. It becomes clear that statistical tests are the beginning and not the end of experimental analyses. Vital in this context is the differentiation between statistically relevant and scientifically meaningful results, which is clearly developed by Thomas Bartz-Beielstein.
This book introduces the new experimentalism in evolutionary computation, providing tools to understand algorithms and programs and their interaction with optimization problems. It develops and applies statistical techniques to analyze and compare modern search heuristics such as evolutionary algorithms and particle swarm optimization. The book bridges the gap between theory and experiment by providing a self-contained experimental methodology and many examples.Part I Basics
Research in Evolutionary Computation
The New Experimentalism
Statistics for Computer Experiments
Optimization Problems
Designs for Computer Experiments
Search Algorithms
Part II Results and Perspectives
Comparison
Understanding Performance
Summary and Outlook