Refinements to search procedures based on simulated annealing for multimodal functions
Aspi Jimmy Havewala · ThinkTech (Texas Tech University) · 1992
1 An oscillatiu~ fnuction of ow" Y(lriahle 1-! •) •) Yariatiou iu the Yalnt>s of 33 3.2 The sin,gle ,•ariahle fun Changes in the :--tep size for .-.in(3.r)+ nJ.~( .r~I n-;iu.u.: the C::\I::\IR algorithm -! 1 3.6 Effect of iucreasiuu_: step size "•ith scaling -!-! 3.S The two YarialJle function .r• y 2 3.9 Choosing direction alternately Yersns S-!.3 \-ariation in optimum ,-alue of the fuuction for De.Joug•s t"•o ,-ariable function 61 -!.-!Effect of inn•easlock..; in the full-scale application of Simulated AmH•nling to fuw•tion ()ptimizntiC~n.This r<'sf'arch problnn ~tudiecl for this t 11<-'sis.1.1 Function Optimization Therc etre m1merous real \Yorld problf'ms.Pach of ,,•hich require le~ defined in a hounded r('u_ion.i.e .. !.!,iYell real Yariahles (.r 1 •.r 2 . .• . . .r,,) and interYals [oi.b,].\Yhere Oi:::; .ri< h,.i = 1.2 ..... n. and a function .f(.r 1 .. r 2 ....•.r 11 ). the problem is to find T = (.r1••~'~• . . . . .1•n) such that .f(Y) is minimum.Algorithms for function minimization haYe been stnclied on dil!;ital computers since the early clays of computing and a number of approaches exist for thf' ta:--k.