Some matrices which are not completely reduced may also be used to give solutions to more general problems with xij > 0. Thus the matrix C(1) of Table 5 with ai > 0 and bj > 0 is a completely reduced matrix for any problem with b2 > a2 + a3 and a4 > b1 + b2 while the C(2) matrix is completely reduced for any problem with b2 > a2 + a3 and b3 < a4 < b1 + b3 . It is useful to note that any negative solution of the problem with fixed specifi- cations corresponds to a non-negative solution of a problem with a4 and bj in- creased by at least the amount of the negative solution. Thus the matrix C(3) of Table 5 is a completely reduced matrix for all problems with specifications ai - 1 + beta , 1, 1, 1; bj = 1, 1 + beta , 1, 1 for all positive integral values of f and for all matrices which reduce to C(3).
Some matrices which are not completely reduced may also be used to give solutions to more general problems with xij > 0. Thus the matrix C(1) of Table 5 with ai > 0 and bj > 0 is a completely reduced matrix for any problem with b2 > a2 + a3 and a4 > b1 + b2 while the C(2) matrix is completely reduced for any problem with b2 > a2 + a3 and b3 < a4 < b1 + b3 . It is useful to note that any negative solution of the problem with fixed specifi- cations corresponds to a non-negative solution of a problem with a4 and bj in- creased by at least the amount of the negative solution. Thus the matrix C(3) of Table 5 is a completely reduced matrix for all problems with specifications ai - 1 + beta , 1, 1, 1; bj = 1, 1 + beta , 1, 1 for all positive integral values of f and for all matrices which reduce to C(3).
การแปล กรุณารอสักครู่..
