Frequently a direct method of solving a problem can be extended to direct
generalizations of the problem. By a direct method we mean one in which the
basic specifications of the problem are used directly in solving the problem without
replacing them, in whole or in part, with auxiliary theorems or criteria and
without using the circuitous approach of transforming an initial feasible solution
to an optimal one. In minimization (maximization) combinatorial problems such
as the transportation problem, the purest direct method consists in writing out
all possible feasible solutions which satisfy (1) and selecting those which satisfy
(2). This method gives a general answer to the problem, in a sense that a circuitous
method does not, and since it is direct it is applicable to many direct
generalizations of the problem which may result from additional specifications
which are similar in form to those of (1). Generally the more direct methods are
subject to more immediate direct generalization since direct methods do not
depend on theorems or criteria which, while they are the complete equivalent of
the conditions of the original problem, are commonly inapplicable to the more
complex ones. Other things being equal, we seek methods which are direct if they
are to be applicable to direct generalizations of the original problem.
But other things are not always equal! The pure direct method of writing out
all possible feasible solutions-those which satisfy (1)-is simply not practical in
most problems. In practice we must resort to methods which are, in part at least,
indirect. The important point is that, if we wish to have methods which are
applicable to direct generalizations of the problem, we should attempt to use a
minimum of equivalence theorems and adixiliary criteria as substitutes for the
stated conditions of the problem.
In considering the nature of many desirable generalizations of the problem,
the minimization condition (2) is usually not fundamentally changed but the
specification conditions (1) are generalized and/or expanded. It appears then
that a proper first step in the order of indirection is in providing an alternative to
the use of (2) while leaving the specifications (1) intact.
Frequently a direct method of solving a problem can be extended to direct
generalizations of the problem. By a direct method we mean one in which the
basic specifications of the problem are used directly in solving the problem without
replacing them, in whole or in part, with auxiliary theorems or criteria and
without using the circuitous approach of transforming an initial feasible solution
to an optimal one. In minimization (maximization) combinatorial problems such
as the transportation problem, the purest direct method consists in writing out
all possible feasible solutions which satisfy (1) and selecting those which satisfy
(2). This method gives a general answer to the problem, in a sense that a circuitous
method does not, and since it is direct it is applicable to many direct
generalizations of the problem which may result from additional specifications
which are similar in form to those of (1). Generally the more direct methods are
subject to more immediate direct generalization since direct methods do not
depend on theorems or criteria which, while they are the complete equivalent of
the conditions of the original problem, are commonly inapplicable to the more
complex ones. Other things being equal, we seek methods which are direct if they
are to be applicable to direct generalizations of the original problem.
But other things are not always equal! The pure direct method of writing out
all possible feasible solutions-those which satisfy (1)-is simply not practical in
most problems. In practice we must resort to methods which are, in part at least,
indirect. The important point is that, if we wish to have methods which are
applicable to direct generalizations of the problem, we should attempt to use a
minimum of equivalence theorems and adixiliary criteria as substitutes for the
stated conditions of the problem.
In considering the nature of many desirable generalizations of the problem,
the minimization condition (2) is usually not fundamentally changed but the
specification conditions (1) are generalized and/or expanded. It appears then
that a proper first step in the order of indirection is in providing an alternative to
the use of (2) while leaving the specifications (1) intact.
การแปล กรุณารอสักครู่..
