4. Optimal solutions

Sun Woong Kim, Steven G. Heeringa and Peter W. Solenberger

Previous | Next

Given the set of L MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaam itaaaa@38D3@ possible arrays in B MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaWefv 3ySLgzgjxyRrxDYbqeguuDJXwAKbIrYf2A0vNCaGqbbiab=fa8cbaa @447C@ , consider the subset B ' ( B ) MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaWefv 3ySLgzgjxyRrxDYbqeguuDJXwAKbIrYf2A0vNCaGqbbiab=fa8cnaa CaaaleqabaGaai4jaaaakiaaykW7caGGOaGaeyOHI0Sae8xaWlKaaG PaVlaacMcaaaa@4D86@ where

p ( B k ) > 0. MathType@MTEF@5@5@+= feaagKart1ev2aaatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaam iCamaabmqabaGaamOqamaaBaaaleaacaWGRbaabeaaaOGaayjkaiaa wMcaaiabg6da+iaaicdacaGGUaaaaa@3EE1@

A solution set to a controlled selection problem A MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaam yqaaaa@38C8@ denoted as

{ ( B k , p ( B k ) ) , B k B ' } MathType@MTEF@5@5@+= feaagKart1ev2aaatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaWaai WaaeaadaqadaqaaiaadkeadaWgaaWcbaGaam4AaaqabaGccaGGSaGa amiCamaabmqabaWaiaiovababKaG4SqaiaiocGaG4m4Aaaqajaiojq 2aG9FaiaiocGaGGmOqaaaaaOGaayjkaiaawMcaaaGaayjkaiaawMca aiaacYcacaaMc8UaaGjcVlaadkeadaWgaaWcbaGaam4AaaqabaGccq GHiiIZtuuDJXwAKzKCHTgD1jharyqr1ngBPrgigjxyRrxDYbacfeGa e8xaWl0aaWbaaSqabeaacaGGNaaaaaGccaGL7bGaayzFaaaaaa@5E61@

is the set of the arrays that have the required positive selection probabilities ( p ( B k ) > 0 ) MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaWaae WaaeaacaWGWbWaaeWabeaacaWGcbWaaSbaaSqaaiaadUgaaeqaaaGc caGLOaGaayzkaaGaeyOpa4JaaGimaaGaayjkaiaawMcaaaaa@3FB9@ . This solution set, or simply a “solution” to the controlled selection problem, is usually obtained by an algorithm to control the constraints in (3.1) through (3.6). As described in the introduction, since Goodman and Kish (1950), many algorithms for obtaining solutions to controlled selection problems have been developed.

Until Groves and Hess (1975) suggested a computer algorithm, most solutions were manually obtained in a process that resembled solving a mathematical puzzle. Furthermore, for most problems it is possible that there is more than one solution set that meets the constraints. Since the 1980s, the computer-intensive controlled selection algorithms using transportation theory, network flow, integer programming, and LP have been developed. These algorithms may depend on highly specialized software or may be programmed to run in major software systems.

However, previous solutions ranging from manual to computer-intensive algorithms have rarely been compared empirically using a standard set of performance criteria. Therefore, we begin with the description of a concept called optimal solution sets, or more simply, optimal solutions.

The controlled selection problem A MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaam yqaaaa@38C8@ is only one array, but there may be many possible arrays in B MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaWefv 3ySLgzgjxyRrxDYbqeguuDJXwAKbIrYf2A0vNCaGqbbiab=fa8cbaa @447C@ . Also, only one array B k MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaam OqamaaBaaaleaacaWGRbaabeaaaaa@39E5@ from any solution to A MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaam yqaaaa@38C8@ is randomly chosen by p ( B k ) MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaam iCamaabmaabaWaiaiovababKaG4SqaiaiocGaG4m4Aaaqajaiojq2a G9FaiaiocGaGGmOqaaaaaOGaayjkaiaawMcaaaaa@4371@ as the basis for choosing the stratified sample. In general then, we might define an optimal solution as that satisfying the following requirements (R1 and R2):

R1. The solution is obtained based on appropriate and objective measurements of the closeness between A MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaam yqaaaa@38C8@ and every single array B k MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaam OqamaaBaaaleaacaWGRbaabeaaaaa@39E5@ in B MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaWefv 3ySLgzgjxyRrxDYbqeguuDJXwAKbIrYf2A0vNCaGqbbiab=fa8cbaa @447C@ .

R2. The solution, as much as possible, maximizes the probabilities of selection over the arrays nearest to A MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaam yqaaaa@38C8@ under such measurements as referenced in R1.

The remainder of this section will address how to specify R1 and R2 for optimal solutions. First, in order to define closeness in R1, a real number d ( B k : A ) MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaam izaiaacIcacaWGcbWaaSbaaSqaaiaadUgaaeqaaOGaaiOoaiaadgea caGGPaaaaa@3DB5@ representing the distance between A MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaam yqaaaa@38C8@ and B k MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaam OqamaaBaaaleaacaWGRbaabeaaaaa@39E5@ , can be considered, where d MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaam izaaaa@38EB@ is a distance function that satisfies the following axioms:

(i)
d ( B k , A ) > 0   if   B k A ;   d ( A , A ) = 0 ; MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaam izamaabmqabaGaamOqamaaBaaaleaacaWGRbaabeaakiaacYcacaWG bbaacaGLOaGaayzkaaGaeyOpa4JaaGimaiaabccacaqGGaGaaeyAai aabAgacaqGGaGaaeiiaiaadkeadaWgaaWcbaGaam4AaaqabaGccqGH GjsUcaWGbbGaai4oaiaabccacaWGKbWaaeWabeaacaWGbbGaaiilai aadgeaaiaawIcacaGLPaaacqGH9aqpcaaIWaGaai4oaaaa@5105@
(ii)
d ( B k , A ) = d ( A , B k ) MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaam izamaabmqabaGaamOqamaaBaaaleaacaWGRbaabeaakiaacYcacaWG bbaacaGLOaGaayzkaaGaeyypa0JaamizamaabmqabaGaamyqaiaacY cacaWGcbWaaSbaaSqaaiaadUgaaeqaaaGccaGLOaGaayzkaaaaaa@44B4@ ;
(iii)
d ( B k , A ) d ( B k , B k ' ) + d ( B k ' , A )  for any  B k ' B . MathType@MTEF@5@5@+= feaagKart1ev2aaatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiqabeqabmaabaabaaGcbaGaam izamaabmqabaGaamOqamaaBaaaleaacaWGRbaabeaakiaacYcacaWG bbaacaGLOaGaayzkaaGaeyizImQaamizamaabmqabaGaamOqamaaBa aaleaacaWGRbaabeaakiaacYcacaWGcbWaa0baaSqaaiaadUgaaeaa caGGNaaaaaGccaGLOaGaayzkaaGaey4kaSIaamizamaabmqabaGaam OqamaaDaaaleaacaWGRbaabaGaai4jaaaakiaacYcacaWGbbaacaGL OaGaayzkaaGaaeiiaiaabAgacaqGVbGaaeOCaiaabccacaqGHbGaae OBaiaabMhacaqGGaGaamOqamaaDaaaleaacaWGRbaabaGaai4jaaaa kiabgIGioprr1ngBPrMrYf2A0vNCaeHbfv3ySLgzGyKCHTgD1jhaiu qacqWFbaVqcqWFUaGlaaa@679D@

Axiom (iii) is termed the triangle inequality axiom. Distance functions satisfying (i), (ii), and (iii) can be defined by using the two-ordered R C -tuples MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaam OuaiaadoeaqaaaaaaaaaWdbiaab2cacaqG0bGaaeyDaiaabchacaqG SbGaaeyzaiaabohaaaa@4020@ ( a 11 , a 12 , , a R C ) MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaWaae WaaeaacaWGHbWaaSbaaSqaaiaaigdacaaIXaaabeaakiaacYcacaWG HbWaaSbaaSqaaiaaigdacaaIYaaabeaakiaacYcacqGHflY1cqGHfl Y1cqGHflY1caGGSaGaamyyamaaBaaaleaacaWGsbGaam4qaaqabaaa kiaawIcacaGLPaaaaaa@4A59@ and ( b 11 k , b 12 k , , b R C k ) MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaWaae WaaeaacaWGIbWaaSbaaSqaaiaaigdacaaIXaGaam4AaaqabaGccaGG SaGaamOyamaaBaaaleaacaaIXaGaaGOmaiaadUgaaeqaaOGaaiilai abgwSixlabgwSixlabgwSixlaacYcacaWGIbWaaSbaaSqaaiaadkfa caWGdbGaam4AaaqabaaakiaawIcacaGLPaaaaaa@4D2C@ for A MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaam yqaaaa@38C8@ and B k MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaam OqamaaBaaaleaacaWGRbaabeaaaaa@39E5@ . We first define the ordinary or Euclidean distance (2-norm distance):

d 2 ( B k , A ) = [ i = 1 R j = 1 C ( b i j k a i j ) 2 ] 1 2 ,     k = 1 , , L . ( 4.1 ) MathType@MTEF@5@5@+= feaagKart1ev2aaatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaam izamaaBaaaleaacaaIYaaabeaakmaabmaabaGaamOqamaaBaaaleaa caWGRbaabeaakiaacYcacaWGbbaacaGLOaGaayzkaaGaeyypa0Zaam WaaKaaafaakmaaqahabaWaaabCaeaadaqadaqaaiaadkgadaWgaaWc baGaamyAaiaadQgacaWGRbaabeaakiabgkHiTiaadggadaWgaaWcba GaamyAaiaadQgaaeqaaaGccaGLOaGaayzkaaWaaWbaaSqabeaacaaI YaaaaaqaaiaadQgacqGH9aqpcaaIXaaabaGaam4qaaqdcqGHris5aa WcbaGaamyAaiabg2da9iaaigdaaeaacaWGsbaaniabggHiLdaajaaq caGLBbGaayzxaaGcdaahaaqcbawabSqaamaalaaabaGaaGymaaqaai acaIgIYaaaaaaakiaacYcacaqGGaGaaeiiaiaabccacaWGRbGaeyyp a0dcbaGaa8xmaiaacYcacqWIMaYscaGGSaGaamitaiaac6cacaaMf8 UaaGzbVlaaywW7caaMf8UaaGzbVlaaywW7caGGOaGaaGinaiaac6ca caaIXaGaaiykaiaaywW7aaa@72EC@

This function is probably the most familiar measure to define the distance between B k MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaam OqamaaBaaaleaacaWGRbaabeaaaaa@39E5@ and A MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaam yqaaaa@38C8@ .

Also, we can define the function called the Chebyshev distance (infinite norm distance):

d ( B k , A ) = max { | b i j k a i j | : i = 1 , , R ,   j = 1 , , C } ,     k = 1 , , L . ( 4.2 ) MathType@MTEF@5@5@+= feaagKart1ev2aaatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaam izamaaBaaaleaacqGHEisPaeqaaOWaaeWabeaacaWGcbWaaSbaaSqa aiaadUgaaeqaaOGaaiilaiaadgeaaiaawIcacaGLPaaacqGH9aqpci GGTbGaaiyyaiaacIhadaGadaqaamaaemaabaGaamOyamaaBaaaleaa caWGPbGaamOAaiaadUgaaeqaaOGaeyOeI0IaamyyamaaBaaaleaaca WGPbGaamOAaaqabaaakiaawEa7caGLiWoacaGG6aGaamyAaiabg2da 9iaaigdacaGGSaGaeSOjGSKaaiilaiaadkfacaGGSaGaaeiiaiaadQ gacqGH9aqpcaaIXaGaaiilaiablAciljaacYcacaWGdbaacaGL7bGa ayzFaaGaaiilaiaabccacaqGGaGaaeiiaiaadUgacqGH9aqpieaaca WFXaGaaiilaiablAciljaacYcacaWGmbGaaiOlaiaaywW7caaMf8Ua aGzbVlaaywW7caaMf8UaaiikaiaaisdacaGGUaGaaGOmaiaacMcaaa a@7328@

These distance functions give rise to distinct distance spaces. Owing to (3.2), for any B k MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaam OqamaaBaaaleaacaWGRbaabeaaaaa@39E5@ , the following holds.

0 d 2 ( B k , A ) < ( R C ) 1 / 2 ( 4.3 ) MathType@MTEF@5@5@+= feaagKart1ev2aaatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaaG imaiabgsMiJkaadsgadaWgaaWcbaGaaGOmaaqabaGcdaqadeqaaiaa dkeadaWgaaWcbaGaam4AaaqabaGccaGGSaGaamyqaaGaayjkaiaawM caaiabgYda8maabmaabaGaamOuaiaadoeaaiaawIcacaGLPaaadaah aaWcbeqaaiaaigdacaGGVaGaaGOmaaaakiaaywW7caaMf8UaaGzbVl aaywW7caaMf8UaaGzbVlaaywW7caaMf8UaaiikaiaaisdacaGGUaGa aG4maiaacMcaaaa@57BB@

and

0 d ( B k , A ) < 1. ( 4.4 ) MathType@MTEF@5@5@+= feaagKart1ev2aaatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0xe9Lq=Jc9 vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKkFr0xfr=x fr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaaGimaiabgs MiJkaadsgadaWgaaWcbaGaeyOhIukabeaakmaabmqabaGaamOqamaa BaaaleaacaWGRbaabeaakiaacYcacaWGbbaacaGLOaGaayzkaaGaey ipaWJaaGymaiaac6cacaaMf8UaaGzbVlaaywW7caaMf8UaaGzbVlaa ywW7caaMf8UaaGzbVlaaywW7caaMf8UaaiikaiaaisdacaGGUaGaaG inaiaacMcaaaa@5556@

For instance, for the 3 × 3 MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaaG 4maiabgEna0kaaiodaaaa@3B93@ array in Problem 2.1 and the 8 × 3 MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaaG ioaiabgEna0kaaiodaaaa@3B98@ array in Problem 2.3, 0 < d 2 ( B k , A ) < 3 MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaacba Gaa8hmaiabgYda8iaadsgadaWgaaWcbaGaaGOmaaqabaGcdaqadaqa aiaadkeadaWgaaWcbaGaam4AaaqabaGccaGGSaGaamyqaaGaayjkai aawMcaaiabgYda8iaa=ndaaaa@423F@ and 0 < d 2 ( B k , A ) < 4 . 9 MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaacba Gaa8hmaiabgYda8iaadsgadaWgaaWcbaGaaGOmaaqabaGcdaqadaqa aiaadkeadaWgaaWcbaGaam4AaaqabaGccaGGSaGaamyqaaGaayjkai aawMcaaiabgYda8iaa=rdacaGGUaGaa8xoaaaa@43AC@ , respectively.

Second, as mentioned in R2, regarding the arrays nearest to A MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaam yqaaaa@38C8@ under such measurements described in R1, consider the set of arrays in B MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaWefv 3ySLgzgjxyRrxDYbqeguuDJXwAKbIrYf2A0vNCaGqbbiab=fa8cbaa @447C@ having the minimum d 2 MathType@MTEF@5@5@+= feaagKart1ev2aaatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaia iMykW7caWGKbWaaSbaaSqaaiaaikdaaeqaaaaa@3C7D@ or d MathType@MTEF@5@5@+= feaagKart1ev2aaatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaam izamaaBaaaleaacqGHEisPaeqaaaaa@3A87@ value from A MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaam yqaaaa@38C8@ . Let B 2 ( B ) MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaWefv 3ySLgzgjxyRrxDYbqeguuDJXwAKbIrYf2A0vNCaGqbbiab=fa8cnaa BaaaleaacaaIYaaabeaakiaacIcacqGHgksZcuWFbaVqgaqbaiaacM caaaa@4A8C@ be the set of the arrays having the minimum d 2 MathType@MTEF@5@5@+= feaagKart1ev2aaatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaia iMykW7caWGKbWaaSbaaSqaaiaaikdaaeqaaaaa@3C7D@ value from A MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaam yqaaaa@38C8@ and B ( B ) MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaWefv 3ySLgzgjxyRrxDYbqeguuDJXwAKbIrYf2A0vNCaGqbbiab=fa8cnaa BaaaleaacqGHEisPaeqaaOGaaiikaiabgAOinlqb=fa8czaafaGaai ykaaaa@4B41@ be the set of the arrays having the minimum d MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaam izamaaBaaaleaacqGHEisPaeqaaaaa@3A88@ value from A MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaam yqaaaa@38C8@ .

Assuming that all possible arrays in B MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaWefv 3ySLgzgjxyRrxDYbqeguuDJXwAKbIrYf2A0vNCaGqbbiab=fa8cbaa @447C@ are known, we define optimum arrays as follows.

Definition. The arrays in B 2 B MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaWefv 3ySLgzgjxyRrxDYbqeguuDJXwAKbIrYf2A0vNCaGqbbiab=fa8cnaa BaaaleaacaaIYaaabeaakiablQIivjab=fa8cnaaBaaaleaacqGHEi sPaeqaaaaa@49F5@ are called the optimum arrays.

Note that in the new algorithm for controlled selection to be described in Section 6, d 2 MathType@MTEF@5@5@+= feaagKart1ev2aaatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaia iMykW7caWGKbWaaSbaaSqaaiaaikdaaeqaaaaa@3C7D@ or d MathType@MTEF@5@5@+= feaagKart1ev2aaatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaia iMykW7caWGKbWaaSbaaSqaaiabg6HiLcqabaaaaa@3D32@ are chosen based on preference. We avoid defining the intersection of B 2 MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaWefv 3ySLgzgjxyRrxDYbqeguuDJXwAKbIrYf2A0vNCaGqbbiab=fa8cnaa BaaaleaacaaIYaaabeaaaaa@4564@ and B MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaWefv 3ySLgzgjxyRrxDYbqeguuDJXwAKbIrYf2A0vNCaGqbbiab=fa8cnaa BaaaleaacqGHEisPaeqaaaaa@4619@ as the optimum arrays because this may exclude the other arrays not in B 2 B MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaWefv 3ySLgzgjxyRrxDYbqeguuDJXwAKbIrYf2A0vNCaGqbbiab=fa8cnaa BaaaleaacaaIYaaabeaakiablMIijjab=fa8cnaaBaaaleaacqGHEi sPaeqaaaaa@49E8@ with the same minimum d 2 MathType@MTEF@5@5@+= feaagKart1ev2aaatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaia iMykW7caWGKbWaaSbaaSqaaiaaikdaaeqaaaaa@3C7D@ ( d MathType@MTEF@5@5@+= feaagKart1ev2aaatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaia iMykW7caWGKbWaaSbaaSqaaiabg6HiLcqabaaaaa@3D32@ ) value. We illustrate below that there may exist a very small number of optimum arrays relative to the number of all possible arrays in B MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaWefv 3ySLgzgjxyRrxDYbqeguuDJXwAKbIrYf2A0vNCaGqbbiab=fa8cbaa @447C@ for any A MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaam yqaaaa@38C8@ . The details of how to find all possible arrays will be described in Section 6 and Section 7.

Illustrations.

For Problem 2.1 through Problem 2.4, it is noted that B 2 B MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaWefv 3ySLgzgjxyRrxDYbqeguuDJXwAKbIrYf2A0vNCaGqbbiab=fa8cnaa BaaaleaacaaIYaaabeaakiabgAOinlab=fa8cnaaBaaaleaacqGHEi sPaeqaaaaa@4AC4@ . Thus, it is possible to use d MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaam izamaaBaaaleaacqGHEisPaeqaaaaa@3A88@ only in illustrating the optimum arrays.

  1. For Problem 2.1, there are six possible arrays satisfying (3.1), (3.2), (3.3), and (3.4). That is, B = { B k , k = 1 , , 6 } MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaWefv 3ySLgzgjxyRrxDYbqeguuDJXwAKbIrYf2A0vNCaGqbbiab=fa8cjab g2da9maacmaabaWaiaiovabajaaqbKaG4SqaiaiocGaGqn4Aaaqcba uajaiojq2aG9FaiaiocGaGimOqaaaakiaacYcacaaMi8UaaGPaVlaa dUgacqGH9aqpcaaIXaGaaiilaiablAciljaacYcacaaI2aaacaGL7b GaayzFaaGaaGjcVlaayIW7aaa@5E18@ , as given in Table 4.1. There exists only one optimum array, B 2 MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaWaia iovabajaaqbKaG4SqaiaiocGaG4GOmaaqcbauajaiojq2aG9Faiaio cGaGimOqaaaaaaa@4130@ , with the minimum value of d = 0.5 MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaam izamaaBaaaleaacqGHEisPaeqaaOGaeyypa0JaaGimaiaac6cacaaI 1aaaaa@3DC3@ .

Table 4.1
3×3 MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqipu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaaG 4maiabgEna0kaaiodaaaa@3B83@ Controlled selection problem, optimum array with d =0.5 MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqipu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaam izamaaBaaaleaacqGHEisPaeqaaOGaeyypa0JaaGimaiaac6cacaaI 1aaaaa@3DB3@ and the other arrays
Table summary
This table displays the results of 3×3 MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqipu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaaG 4maiabgEna0kaaiodaaaa@3B83@ Controlled selection problem. The information is grouped by A MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqk0Jf9crFfpeea0xh9v8qiW7rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaam yqaaaa@3AEB@ , B 1 MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqk0Jf9crFfpeea0xh9v8qiW7rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaWaia iovabajaaqbKaG4SqaiaiocGaG4Gymaaqcbauajaiojq2aG9Faiaio cGaGimOqaaaaaaa@4352@ , B 2 MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqk0Jf9crFfpeea0xh9v8qiW7rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaWaia iovabajaaqbKaG4SqaiaiocGaG4GOmaaqcbauajaiojq2aG9Faiaio cGaGimOqaaaaaaa@4353@ , B 3 MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqk0Jf9crFfpeea0xh9v8qiW7rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaWaia iovabajaaqbKaG4SqaiaiocGaG4G4maaqcbauajaiojq2aG9Faiaio cGaGimOqaaaaaaa@4354@ , B 5 MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqk0Jf9crFfpeea0xh9v8qiW7rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaWaia iovabajaaqbKaG4SqaiaiocGaG4Gynaaqcbauajaiojq2aG9Faiaio cGaGimOqaaaaaaa@4356@ , B 4 MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqk0Jf9crFfpeea0xh9v8qiW7rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaWaia iovabajaaqbKaG4SqaiaiocGaG4Ginaaqcbauajaiojq2aG9Faiaio cGaGimOqaaaaaaa@4355@ , B 6 MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqk0Jf9crFfpeea0xh9v8qiW7rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaWaia iovabajaaqbKaG4SqaiaiocGaG4GOnaaqcbauajaiojq2aG9Faiaio cGaGimOqaaaaaaa@4357@ (appearing as column headers).
A MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqk0Jf9crFfpeea0xh9v8qiW7rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaam yqaaaa@3AEB@ B 1 MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqk0Jf9crFfpeea0xh9v8qiW7rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaWaia iovabajaaqbKaG4SqaiaiocGaG4Gymaaqcbauajaiojq2aG9Faiaio cGaGimOqaaaaaaa@4352@ B 2 MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqk0Jf9crFfpeea0xh9v8qiW7rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaWaia iovabajaaqbKaG4SqaiaiocGaG4GOmaaqcbauajaiojq2aG9Faiaio cGaGimOqaaaaaaa@4353@ B 3 MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqk0Jf9crFfpeea0xh9v8qiW7rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaWaia iovabajaaqbKaG4SqaiaiocGaG4G4maaqcbauajaiojq2aG9Faiaio cGaGimOqaaaaaaa@4354@ B 4 MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqk0Jf9crFfpeea0xh9v8qiW7rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaWaia iovabajaaqbKaG4SqaiaiocGaG4Ginaaqcbauajaiojq2aG9Faiaio cGaGimOqaaaaaaa@4355@ B 5 MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqk0Jf9crFfpeea0xh9v8qiW7rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaWaia iovabajaaqbKaG4SqaiaiocGaG4Gynaaqcbauajaiojq2aG9Faiaio cGaGimOqaaaaaaa@4356@ B 6 MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqk0Jf9crFfpeea0xh9v8qiW7rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaWaia iovabajaaqbKaG4SqaiaiocGaG4GOnaaqcbauajaiojq2aG9Faiaio cGaGimOqaaaaaaa@4357@
0.8 0.5 0.7 0.7 0.8 0.5 0.5 0.7 0.8 MathType@MTEF@5@5@+= feaagKart1ev2aaatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbGeaGqiVu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0xe9Lq=Jc9 vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKkFr0xfr=x fr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaqbaeqabmWaaa qaaabaaaaaaaaapeGaaGimaiaac6cacaaI4aaapaqaa8qacaaIWaGa aiOlaiaaiwdaa8aabaWdbiaaicdacaGGUaGaaG4naaWdaeaapeGaaG imaiaac6cacaaI3aaapaqaa8qacaaIWaGaaiOlaiaaiIdaa8aabaWd biaaicdacaGGUaGaaGynaaWdaeaapeGaaGimaiaac6cacaaI1aaapa qaa8qacaaIWaGaaiOlaiaaiEdaa8aabaWdbiaaicdacaGGUaGaaGio aaaaaaa@4AB7@ 0 1 1 1 0 1 1 1 0 MathType@MTEF@5@5@+= feaagKart1ev2aaatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbGeaGqiVu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0xe9Lq=Jc9 vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKkFr0xfr=x fr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbiqaaGjbfaqabe Wadaaabaaeaaaaaaaaa8qacaaIWaaapaqaaiaaigdaaeaacaaIXaaa baGaaGymaaqaa8qacaaIWaaapaqaaiaaigdaaeaacaaIXaaabaGaaG ymaaqaa8qacaaIWaaaaaaa@3D9A@ 1 0 1 1 1 0 0 1 1 MathType@MTEF@5@5@+= feaagKart1ev2aaatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbGeaGqiVu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0xe9Lq=Jc9 vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKkFr0xfr=x fr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbiqaaGjbfaqabe WadaaabaGaaGymaaqaaiaaicdaaeaacaaIXaaabaGaaGymaaqaaiaa igdaaeaacaaIWaaabaGaaGimaaqaaiaaigdaaeaacaaIXaaaaaaa@3D3C@ 1 1 0 0 1 1 1 0 1 MathType@MTEF@5@5@+= feaagKart1ev2aaatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbGeaGqiVu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0xe9Lq=Jc9 vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKkFr0xfr=x fr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbiqaaGjbfaqabe WadaaabaGaaGymaaqaaiaaigdaaeaacaaIWaaabaGaaGimaaqaaiaa igdaaeaacaaIXaaabaGaaGymaaqaaiaaicdaaeaacaaIXaaaaaaa@3D3C@ 0 1 1 1 1 0 1 0 1 MathType@MTEF@5@5@+= feaagKart1ev2aaatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbGeaGqiVu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0xe9Lq=Jc9 vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKkFr0xfr=x fr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbiqaaGjbfaqabe WadaaabaGaaGimaaqaaiaaigdaaeaacaaIXaaabaGaaGymaaqaaiaa igdaaeaacaaIWaaabaGaaGymaaqaaiaaicdaaeaacaaIXaaaaaaa@3D3C@ 1 0 1 0 1 1 1 1 0 MathType@MTEF@5@5@+= feaagKart1ev2aaatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbGeaGqiVu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0xe9Lq=Jc9 vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKkFr0xfr=x fr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbiqaaGjbfaqabe WadaaabaGaaGymaaqaaiaaicdaaeaacaaIXaaabaGaaGimaaqaaiaa igdaaeaacaaIXaaabaGaaGymaaqaaiaaigdaaeaacaaIWaaaaaaa@3D3C@ 1 1 0 1 0 1 0 1 1 MathType@MTEF@5@5@+= feaagKart1ev2aaatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbGeaGqiVu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0xe9Lq=Jc9 vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKkFr0xfr=x fr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbiqaaGjbfaqabe WadaaabaGaaGymaaqaaiaaigdaaeaacaaIWaaabaGaaGymaaqaaiaa icdaaeaacaaIXaaabaGaaGimaaqaaiaaigdaaeaacaaIXaaaaaaa@3D3C@
d MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqk0Jf9crFfpeea0xh9v8qiW7rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaam izamaaBaaaleaacqGHEisPaeqaaaaa@3CAB@ 0.8 0.5 0.7 0.8 0.8 0.8
  1. Problem 2.2 has 30 possible arrays, and there are three optimum arrays, shown in Table 4.2

Table 4.2
4×4 MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqipu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaaG inaiabgEna0kaaisdaaaa@3B85@ Optimum arrays with d =0.6 MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqipu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaam izamaaBaaaleaacqGHEisPaeqaaOGaeyypa0JaaGimaiaac6cacaaI 2aaaaa@3DB4@

0 0 1 1 1 1 0 0 0 0 1 1 1 1 0 0 MathType@MTEF@5@5@+= feaagKart1ev2aaatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbGeaGqiVu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0xe9Lq=Jc9 vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKkFr0xfr=x fr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbiqaaGjbfaqabe abeaaaaaqaaabaaaaaaaaapeGaaGimaaWdaeaapeGaaGimaaWdaeaa caaIXaaabaGaaGymaaqaaiaaigdaaeaacaaIXaaabaGaaGimaaqaai aaicdaaeaacaaIWaaabaGaaGimaaqaaiaaigdaaeaacaaIXaaabaGa aGymaaqaaiaaigdaaeaacaaIWaaabaGaaGimaaaaaaa@42AB@
0 1 1 0 1 0 0 1 0 0 1 1 1 1 0 0 MathType@MTEF@5@5@+= feaagKart1ev2aaatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbGeaGqiVu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0xe9Lq=Jc9 vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKkFr0xfr=x fr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbiqaaGjbfaqabe abeaaaaaqaaiaaicdaaeaacaaIXaaabaGaaGymaaqaaiaaicdaaeaa caaIXaaabaGaaGimaaqaaiaaicdaaeaacaaIXaaabaGaaGimaaqaai aaicdaaeaacaaIXaaabaGaaGymaaqaaiaaigdaaeaacaaIXaaabaGa aGimaaqaaiaaicdaaaaaaa@425D@
0 1 1 0 1 0 1 0 1 0 0 1 0 1 0 1 MathType@MTEF@5@5@+= feaagKart1ev2aaatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbGeaGqiVu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0xe9Lq=Jc9 vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKkFr0xfr=x fr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbiqaaGjbfaqabe abeaaaaaqaaiaaicdaaeaacaaIXaaabaGaaGymaaqaaiaaicdaaeaa caaIXaaabaGaaGimaaqaaiaaigdaaeaacaaIWaaabaGaaGymaaqaai aaicdaaeaacaaIWaaabaGaaGymaaqaaiaaicdaaeaacaaIXaaabaGa aGimaaqaaiaaigdaaaaaaa@425D@
  1. Problem 2.3 has 141 possible arrays. There are six optimum arrays, where each array has the same d = 0.6 MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaam izamaaBaaaleaacqGHEisPaeqaaOGaeyypa0JaaGimaiaac6cacaaI 2aaaaa@3DC4@ . One of them is given in Table 4.3.

Table 4.3
One of six optimum arrays with d =0.6 MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqipu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaam izamaaBaaaleaacqGHEisPaeqaaOGaeyypa0JaaGimaiaac6cacaaI 2aaaaa@3DB4@

1 2 0 1 0 1 0 0 0 1 1 0 1 1 0 0 0 1 0 0 0 0 0 0 MathType@MTEF@5@5@+= feaagKart1ev2aaatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbGeaGqiVu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0xe9Lq=Jc9 vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKkFr0xfr=x fr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbiqaaGjbfaqabe acdaaaaaqaaiaaigdaaeaacaaIYaaabaGaaGimaaqaaiaaigdaaeaa caaIWaaabaGaaGymaaqaaiaaicdaaeaacaaIWaaabaGaaGimaaqaai aaigdaaeaacaaIXaaabaGaaGimaaqaaiaaigdaaeaacaaIXaaabaGa aGimaaqaaiaaicdaaeaacaaIWaaabaGaaGymaaqaaiaaicdaaeaaca aIWaaabaGaaGimaaqaaiaaicdaaeaacaaIWaaabaGaaGimaaaaaaa@483A@
  1. There are 159 possible arrays for Problem 2.4, and there is only one optimum array given in Table 4.4.

Table 4.4
5×5 MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqipu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaaG ynaiabgEna0kaaiwdaaaa@3B87@ Optimum array with d =0.517 MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqipu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaam izamaaBaaaleaacqGHEisPaeqaaOGaeyypa0JaaGimaiaac6cacaaI 1aGaaGymaiaaiEdaaaa@3F2F@

2 3 1 0 0 2 1 1 1 1 0 2 2 2 1 1 0 1 3 3 1 0 2 2 5 MathType@MTEF@5@5@+= feaagKart1ev2aaatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbGeaGqiVu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0xe9Lq=Jc9 vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKkFr0xfr=x fr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbiqaaGjbfaqabe qbfaaaaaqaaiaaikdaaeaacaaIZaaabaGaaGymaaqaaiaaicdaaeaa caaIWaaabaGaaGOmaaqaaiaaigdaaeaacaaIXaaabaGaaGymaaqaai aaigdaaeaacaaIWaaabaGaaGOmaaqaaiaaikdaaeaacaaIYaaabaGa aGymaaqaaiaaigdaaeaacaaIWaaabaGaaGymaaqaaiaaiodaaeaaca aIZaaabaGaaGymaaqaaiaaicdaaeaacaaIYaaabaGaaGOmaaqaaiaa iwdaaaaaaa@490F@

Accordingly, based on the definition of optimum arrays as well as d 2 MathType@MTEF@5@5@+= feaagKart1ev2aaatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaia iMykW7caWGKbWaaSbaaSqaaiaaikdaaeqaaaaa@3C7D@ and d MathType@MTEF@5@5@+= feaagKart1ev2aaatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaam izamaaBaaaleaacqGHEisPaeqaaaaa@3A87@ satisfying the axioms (i), (ii), and (iii), we suggest the following specifications (S1 and S2) of R1 and R2 of optimal solutions:

S1. The solution is based on the values of the distance d 2 ( d ) MathType@MTEF@5@5@+= feaagKart1ev2aaatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaia iMykW7caWGKbWaaSbaaSqaaiaaikdaaeqaaOWaaeWaaeaacaWGKbWa aSbaaSqaaiabg6HiLcqabaaakiaawIcacaGLPaaaaaa@40A0@ between A MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaam yqaaaa@38C8@ and every single array B k MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaam OqamaaBaaaleaacaWGRbaabeaaaaa@39E5@ in B MathType@MTEF@5@5@+= feaagKart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLn hiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr 4rNCHbcvPDwzYbGeaGqiFu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0x e9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKk Fr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaWefv 3ySLgzgjxyRrxDYbqeguuDJXwAKbIrYf2A0vNCaGqbbiab=fa8cbaa @447C@ .

S2. The solution maximizes the probabilities of selection of optimum arrays.

S1 and S2 will be the rudiments of a new algorithm presented in Section 6, and in the next section we turn into the discussion on the previous algorithms from the viewpoint of optimal solutions.

Previous | Next

Report a problem on this page

Is something not working? Is there information outdated? Can't find what you're looking for?

Please contact us and let us know how we can help you.

Privacy notice

Date modified: