Scaling of run "time" with dimension
Average running time (aRT in number of f−evaluations as log10 value), divided by dimension for target function value 10^{−8} versus dimension. Slanted grid lines indicate quadratic scaling with the dimension. Different symbols correspond to different algorithms given in the legend of f1 and f24. Light symbols give the maximum number of function evaluations from the longest trial divided by dimension. Black stars indicate a statistically better result compared to all other algorithms with p < 0.01 and Bonferroni correction number of dimensions (six). Legend: ○: ACOR liao noiseless, ♢: BIPOPaCMA loshchilov noiseless, ☆: BIPOPsaACM loshchilov noiseless, ▼: CMAES posik noiseless, ⬡: CMA brockhoff noiseless, △: CMAa brockhoff noiseless, ⬠: CMAm brockhoff noiseless, rightY: CMAma brockhoff noiseless, downY: CMAmah brockhoff noiseless, ◇: CMAmh brockhoff noiseless, ◁: DBRCGA chuang noiseless, ▼: DE-AUTO voglis noiseless, ☆: DE-BFGS voglis noiseless, ◻: DE-ROLL voglis noiseless, ♢: DE-SIMPLEX voglis noiseless, △: DEAE posik noiseless, ◁: DE posik noiseless, ⬡: DEb posik noiseless, ⬠: DEctpb posik noiseless, ⬣: IPOPsaACM loshchilov noiseless, downY: JADE posik noiseless, upY: JADEb posik noiseless, leftY: JADEctpb posik noiseless, rightY: MVDE melo noiseless, ◇: NBIPOPaCMA loshchilov noiseless, ♢: NIPOPaCMA loshchilov noiseless, ☆: PSO-BFGS voglis noiseless, ▼: SNES schaul noiseless, ⬡: xNES schaul noiseless, △: xNESas schaul noiseless