正在加载图片...
minf(x)x=(xP,xc)∈Rn s.t.9,(x)≤0j=1,m xxi=1,n here f(x)and g;(x)(j=1,m)are permitted to be linear or nonlinear functions of x.=(x,x2.,)ER is a subset of discrete variables (the integer variablcs are considered as a special case of discrete ones)x=(, x.)ER is a subset of continuous variables.However,according to the fact that either of Ror Rc is empty,or both are not empty,there is an all-conti- nuous,an all-discrete,or a mixed discrete optimization problems,respectively. From the point of view of engincering design,variables that are inherently continuous wilk be allowed to discretize arbitrarily into k,value,so the discrete intervals are: e:=(x'-x)/(k,-1)j=p+1,n (2) where e.is called quasi-discrete increment,its values are determined by the requirement of engineering design,and so the problems (1)could be considered as an all-discrete problems. The discrete space is defined as a set of all discrete values,i.e. RD={xi}xs∈R (3) and so the feasible design region is defincd as D={xg,(x)≤0V,∈RPCR"} (4) The difference between two neighbouring discrete values with any variable axis direction is defined as the increment,i.e. PI..-increment f=x1-i j=1,k,i=1,n (5) Minus-increment4=x- general☑il卡l☑l,but for quasi-discrete increment=ldl=et, Assuming that x ERCR",the set =(i i=1,p1 (6) j=p+1,n) is defined as the unit neighborhood of the discrete point x,and then the set NC(x)=xUN(x)ne i=1,n) (7) is defined as the coordinate neighborhood of the discrete point x. The discrete optimum x'of problem (1)have to enclose this condition that could be formulated by the following active constraints: 341二 , 二 亡 任 ’ 一 。 , 劣 簇 , ,, 、 二 ,簇 犷 , , 气 主 二 夕 , , 左 劣 ,, , , … , , 任 , , , 尸 , … 〔 、 · , , , 一 , 一 , , 。 , 舟 ‘ , 口户尹 。 。 于 ‘ 一 二 于 八掩 ‘ 一 , 。 ‘ 一 , , 一 , 一 二 ‘ , 。 、 任 “ 】夕, 簇 ,任 ” , 。 ‘ 一 」 ‘ , 一 劣 ‘ , , , , ” 一 了 二 劣 ‘ ,一 一 劣 ‘ , 」李 午 」了 , 一 亡 专 」下 ‘ 。 任 二 ’ , “ , 劣 」下, ‘ , ‘ 」 一 £ , 劣 一 劣 君 了 , , 月 二 , 二 二 自 , 川 劣 ’
<<向上翻页向下翻页>>
©2008-现在 cucdc.com 高等教育资讯网 版权所有