Steady-State Procedure forReal-Parameter Optimization
Santosh Kumar Tiwari · 2005
Despite theexistence ofanumberofproce- dures forreal-parameter optimization using evolution- aryalgorithms, there isstill aneedofasystematic and unbiased comparison ofdifferent approaches onacare- fully chosen setoftest problems. Inthis paper, wede- velop asteady-state, population-based optimization al- gorithm whichallows themainsearch principles tobe independently designed. Thealgorithm sodeveloped is applied toasetof25testproblems andresults on10 and30dimensions arepresented. Although thepro- posed procedure cannot find theexact optimum within thespecified numberoffunction evaluations, inmost problems, thealgorithm showssteady progress towards theoptimum. Moreover, itisalso observed that theper- formance ofthealgorithm doesnotgetaffected bythe rotation ofthefunctions, discontinuity andembedded noise infunction description.