04.02.2015 Views

Stochastic Programming - Index of

Stochastic Programming - Index of

Stochastic Programming - Index of

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

6 STOCHASTIC PROGRAMMING<br />

We observed from Table 1 that the expected Net Present Value (NPV)<br />

<strong>of</strong> Decision 4, i.e. the decision to develop Lot 2 and build a plant, equals<br />

30. Standard theory tells us to invest if a project has a positive NPV, since<br />

that means the project is pr<strong>of</strong>itable. And, indeed, Decision 4 represents an<br />

investment which is pr<strong>of</strong>itable in terms <strong>of</strong> expected pr<strong>of</strong>its. But as we have<br />

observed, Decision 3 is better, and it is not possible to make both decisions;<br />

they exclude each other. The expected NPV for Decision 3 is 195. The<br />

difference <strong>of</strong> 165 is the value <strong>of</strong> an option, namely the option not to build<br />

the plant. Or to put it in a different wording: If your only possibilities were to<br />

develop Lot 2 and build the plant at the same time, or do nothing, and you<br />

were asked how much you were willing to pay in order to be allowed to delay<br />

the building <strong>of</strong> the plant (at the 10% penalty) the answer is at most 165.<br />

Another possible setting is to assume that the right to develop Lot 2 and<br />

build the plant is for sale. This right can be seen as an option. This option is<br />

worth 195 in the setting where delayed construction <strong>of</strong> the plant is allowed.<br />

(If delays were not allowed, the right to develop and build would be worth 30,<br />

but that is not an option.)<br />

So what is it that gives an option a value Its value stems from the right<br />

to do something in the future under certain circumstances, but to drop it in<br />

others if you so wish. And, even more importantly, to evaluate an option you<br />

must model explicitly the future decisions. This is true in our simple model,<br />

but it is equally true in any complex option model. It is not enough to describe<br />

a stochastic future, this stochastic future must contain decisions.<br />

So what are the important aspect <strong>of</strong> randomness We may conclude that<br />

there are at least three (all related <strong>of</strong> course).<br />

1. Randomness is needed to obtain a correct evaluation <strong>of</strong> the future income<br />

and costs, i.e. to evaluate the objective.<br />

2. Flexibility only has value (and meaning) in a setting <strong>of</strong> randomness.<br />

3. Only by explicitly evaluating future decisions can decisions containing<br />

flexibility (options) be correctly valued.

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!