the final choice. For example, in Fig. 14.3, all alternatives X k which are on the
noninferior part BCDE of the feasible boundary curve are not dominated; i.e., no
other alternatives having equal or higher values for any of the goals exist. These
nondominated alternatives are often named Pareto optimal or Pareto admissible. As
shown on Fig. 14.3, there are often several Pareto optimal or efficient alternatives.
The satisficing method also reduces the number of alternative solutions. For each
objective, the decision maker gives the minimum goal values acceptable to him.
Alternative schemes not meeting these minimum targets are discarded. The minimum values to be achieved for one or more goals are raised in the remaining
alternative solutions which are again screened. This procedure is carried out iteratively until a single scheme remains. Here again, satisficing does not need numerical
values for each goal. As minimum values have to be satisfied, alternatives with some
high goal values do not receive any special credit, either.
Lexicography involves the ranking of alternatives based on a comparison of
objective values for each objective, one at a time, and in order of their importance to
the decision maker. Thus the decision maker must specify which objective is most
important and then compare the values of this objective for all alternatives. If for this
most important objective one alternative has a higher objective value than any of the
other alternatives, then it is chosen and the decision process ends. If there is more
than one alternative having the highest value of that most important objective, then
these and only these alternatives are considered in the next comparison based on the
second-most-important objective. The process continues in this lexicographic fashion until either a single alternative emerges or all objectives have been considered.
Although simple and not requiring much information, each of these methods has
a number of limitations, which may be overcome by more complex assessment
techniques, this time needing more information and assumptions.
14.7.2 STEM, an Iterative Procedure
Benayoun et al. (1971) have also suggested involving the decision maker throughout
the mathematical programming framework using a progressive iteration and exploration approach that ‘explains’ to the decision maker how to distinguish between
what he contemplates as good solutions and important goals. The final solution
chosen by the decision maker characterises a best concession among conflicting
goals. If preferences do not vary during the process, quite a small number of
iterations are needed for the decision maker to achieve the best compromise. The
authors have labelled their approach STEM (step method).
The decision maker is initially assumed incapable of defining the relative significance of the separate objectives F j (X). Throughout a certain number of calculations
and decision-making iterations, the decision maker is taught. This essentially
involves a dialogue between the decision maker and the analyst. Each iteration is
followed by a decision-making phase, during which the decision maker scrutinises
the results from the calculation phase and cultivates new perceptions and statistics
14.7 Plan Selection: The Identification of Politically Feasible Alternatives
401
noninferior part BCDE of the feasible boundary curve are not dominated; i.e., no
other alternatives having equal or higher values for any of the goals exist. These
nondominated alternatives are often named Pareto optimal or Pareto admissible. As
shown on Fig. 14.3, there are often several Pareto optimal or efficient alternatives.
The satisficing method also reduces the number of alternative solutions. For each
objective, the decision maker gives the minimum goal values acceptable to him.
Alternative schemes not meeting these minimum targets are discarded. The minimum values to be achieved for one or more goals are raised in the remaining
alternative solutions which are again screened. This procedure is carried out iteratively until a single scheme remains. Here again, satisficing does not need numerical
values for each goal. As minimum values have to be satisfied, alternatives with some
high goal values do not receive any special credit, either.
Lexicography involves the ranking of alternatives based on a comparison of
objective values for each objective, one at a time, and in order of their importance to
the decision maker. Thus the decision maker must specify which objective is most
important and then compare the values of this objective for all alternatives. If for this
most important objective one alternative has a higher objective value than any of the
other alternatives, then it is chosen and the decision process ends. If there is more
than one alternative having the highest value of that most important objective, then
these and only these alternatives are considered in the next comparison based on the
second-most-important objective. The process continues in this lexicographic fashion until either a single alternative emerges or all objectives have been considered.
Although simple and not requiring much information, each of these methods has
a number of limitations, which may be overcome by more complex assessment
techniques, this time needing more information and assumptions.
14.7.2 STEM, an Iterative Procedure
Benayoun et al. (1971) have also suggested involving the decision maker throughout
the mathematical programming framework using a progressive iteration and exploration approach that ‘explains’ to the decision maker how to distinguish between
what he contemplates as good solutions and important goals. The final solution
chosen by the decision maker characterises a best concession among conflicting
goals. If preferences do not vary during the process, quite a small number of
iterations are needed for the decision maker to achieve the best compromise. The
authors have labelled their approach STEM (step method).
The decision maker is initially assumed incapable of defining the relative significance of the separate objectives F j (X). Throughout a certain number of calculations
and decision-making iterations, the decision maker is taught. This essentially
involves a dialogue between the decision maker and the analyst. Each iteration is
followed by a decision-making phase, during which the decision maker scrutinises
the results from the calculation phase and cultivates new perceptions and statistics
14.7 Plan Selection: The Identification of Politically Feasible Alternatives
401