Skip to main content
zenodoopen

Patterns of convergence and bound constraint violation in Differential Evolution on SBOX-COST benchmarking suite

<p>The data sets contain information collected during runs of :</p> <ul> <li>DE/rand/1/bin (D=20, N=50, F=0.5, CR=0.5);</li> <li>L-SHADE(D=20, N_init = 18*D)</li> </ul> <p>on all 24 test functions from instances 1-5 and 101-110 from the SBOX-COST and BBOB test suites.</p> <ul> <li>For the <strong>analysis with&nbsp;fixed budget</strong>, the number of independent runs was 30 and the&nbsp;stopping condition NFE (number of function evaluations) was 5000*D.</li> <li>For the <strong>convergence analysis</strong>, 5 independent runs were used for instances 1-5 of the 24 test functions and there was a combined stopping condition consisting of&nbsp;NFE=10000*D, population size larger than 10, variance of the population greater than 10^(-15) and the difference between the global best optimum and known best greater than 10^(-8).</li> </ul> <p>&nbsp;</p> <p>&nbsp;</p>

ShareScore

32/100

Overall dataset sharing score

Score breakdown

These five areas show where the dataset supports — or may limit — practical reuse.

Stewardship
4
Harmonization
4
Access
16
Reuse readiness
8
Engagement
0