|
|
|
EXCEL@EAs Beta Version 1.0 (Reed, P. and C. Coronado 2004)
EXCEL@EAs Beta Version 1.0 is an Microsoft EXCEL-based educational software developed to help undergraduate and graduate students study the key issues controlling the efficiency and reliability of evolutionary algorithms. The software uses Microsoft Excel Visual Basic to provide students with an easy-to-use compiler environment where they can directly manipulate four evolutionary algorithm source codes or experiment with the algorithms using simple spreadsheet-based graphical user interfaces.
Downloadable versions of EXCEL@EAs Beta Version 1.0
-
EXCEL@EAs Beta Version 1.0 (Reed, P. and C. Coronado 2004) (download, 523 KB)
-
Interactive Video Tutorial by C. Coronado 2004 (download, 49.4 MB)
Epsilon-NSGAII Version 3.01 Serial Driver (Kollat, J., Tang, Y., and Reed, P. 2006)
This version of the ε-NSGAII is described in the following papers [1, 2]. It would be appreciated that use of this code is properly cited and that users please share subsequent copies of publications generated using the code for our reference. The ε-NSGAII builds on its parent algorithm, the NSGAII developed by Dr. Kalyanmoy Deb by adding ε-dominance archiving and adaptive population sizing to minimize the need for extensive parameter calibration. The concept of ε-dominance allows the user to specify the precision with which they want to quantify each objective in a multi-objective problem. The ε-NSGAII uses a series of “connected runs” where small populations are initially exploited to pre-condition search and automatically adapt population size commensurate with problem difficulty. As the search progresses, the population size is automatically adapted based on the number of ε-nondominated solutions that the algorithm has found. Epsilon-non-dominated solutions found after each generation are stored in an archive and subsequently used to direct the search. Theoretically, this approach allows the algorithm’s population size to increase or decrease, and in the limit when the ε-dominance archive size stabilizes, the ε-NSGAII’s “connected runs” are equivalent to a diversity-based EA search enhancement termed “time continuation”.
Code Download and Related Publications