文件列表:

CBBO\Ackley.m (1727, 2007-01-29)
CBBO\BBO.m (8427, 2013-03-07)
CBBO\CBBO11_20.m (8628, 2014-07-22)
CBBO\CBBO1_10.m (8619, 2014-07-22)
CBBO\CBBO21_30.m (8533, 2014-07-22)
CBBO\CBBO31_40.m (8552, 2014-07-22)
CBBO\CBBO41_50.m (8629, 2014-07-22)
CBBO\chaos.m (1694, 2014-07-22)
CBBO\ClearDups.m (743, 2007-01-12)
CBBO\ComputeAveCost.m (443, 2009-02-13)
CBBO\Conclude.m (914, 2007-01-29)
CBBO\CountUnique.m (925, 2008-10-20)
CBBO\Fletcher.m (2285, 2007-01-29)
CBBO\Griewank.m (1623, 2008-12-02)
CBBO\Init.m (1581, 2013-03-07)
CBBO\main.m (4060, 2014-07-22)
CBBO\Monte.m (2940, 2008-03-01)
CBBO\Penalty1.m (2116, 2007-02-13)
CBBO\Penalty2.m (2120, 2007-02-13)
CBBO\PopSort.m (458, 2008-10-22)
CBBO\Quartic.m (1634, 2007-02-13)
CBBO\Rastrigin.m (1744, 2009-02-13)
CBBO\Rosenbrock.m (1866, 2007-01-29)
CBBO\Schwefel.m (1719, 2007-01-29)
CBBO\TransitionMonte.m (5950, 2009-02-13)
CBBO\TransitionMonteAll.m (282, 2009-02-13)
license.txt (1508, 2014-07-22)

A Probabilistic Analysis of a Simplified Biogeography-Based Optimization Algorithm by Dan Simon February 13, 2009 This zip file was downloaded from http://academic.csuohio.edu/simond/bbo/simplified. The files in this zip file were used to obtain the results in the paper mentioned above. The theoretical results in Table 1 in the paper were obtained by running the MATLAB program >> TransitionMonteAll; This program takes about 20 or 30 minutes to run on a 2 GHz PC. The experimental results in Tables 1 and 2 were obtained by running the program >> SimpleBBOMonteAll; This program takes about 30 minutes to run on a 2 GHz PC. The main reason is takes so long is because after every generation it checks if the population is uniform. If you just want to check the results in the paper, then you can go into SimpleBBOMonte.m and change PopSize from [4 16 ***] to [4 16]. Then you won't get the population-size-*** results, but you will get the other results, and it will take about an order of magnitude less time to run. If you just want to run the simplified BBO algorithm on your own optimization problem, you can run >> SimpleBBO(@ProbFunction, DisplayFlag, PopSize, GenFlag, RandSeed); The inputs are: @ProbFunction is the address of the problem function. You can create your own problem function by using the problem functions in this zip file as a template (see Rastrigin.m, Schwefel226.m, Sphere.m, Step.m, and Griewank.m). DisplayFlag is true or false, depending on if you want to display progress status during the algorithm execution. PopSize is the population size. I used population sizes of 4, 16, and *** in the paper. GenFlag is true or false, depending on if you want to automatically terminate the algorithm when the population reaches uniformity. RandSeed is the random number generator seed. You can set this if you want the initial population to be the same every time you run the program. If you don't pass a value for RandSeed, then the clock is used as the random number generator seed. Note that your results will vary somewhat from the paper, because the results are based on a statistical average. The simplified BBO does not really perform very well compared with the standard BBO. This is because it is so simple. The purpose of this work was not to improve BBO, but to simplify BBO so that tools could be developed to analyze it.

近期下载者：

相关文件：

评论：[我要评论] [举报此文件]

收藏者：