ABSTRACT
The standard 3 the figures for the Row maxima column. But, again, in so simple an example, it adds to the interest while causing no difficulty. As it is regret-at-potential-losses we now seek to minimize, we glance along
the rows and pick out the largest potential loss that could arise from each option in turn. The largest figure along row a1 is 10,000. It is therefore placed opposite a1 under the third column containing the row maxima. For row a2, the largest potential loss figure is 140,000, and it is entered accordingly opposite a2 and in the third column. For row a3, the largest potential loss figure is 310,000, and this is shown opposite a3 in the third column. Of these largest potential losses from choosing a1, a2 or a3, the decision maker
chooses the smallest, which is 10,000, corresponding to option a1. Accordingly, the figure of 10,000 is entered in the fourth column of Table 39.2. By choosing option a1, he can be sure of one thing: that whichever event occurs, his potential loss – that is, the additional gain he might, in that event, have obtained had he instead selected one of the other options – can be no greater than 10,000. For clearly, if instead he chooses a2, the potential loss he may suffer is 140,000. While if he chooses a3, the potential loss he may suffer is 310,000.1
As a further illustration of the minimax-regret method, Table 39.3 is constructed from the primary data given in Table 38.3. For the first column, below b1 the best choice is a2, as it would entail the least cost 11.3. If option a1 were chosen instead,
QUAH: “CHAP39” — 2007/1/25 — 08:00 — PAGE 215 — #4
would be 1.7 3 be 12.8, and the loss of potential savings would be 1.5 (12.8 minus 11.3). The figures in the next two columns are obtained in the same way. In the fourth column, we put the row maxima. From these, the smallest potential loss, 0.5 corresponding to option a2, is chosen and entered in the fifth column.
