Angaben zur Quelle [Bearbeiten]
| Autor | John R. Schuyler |
| Titel | Credible Projections Now Require Decision Analysis |
| Zeitschrift | Cost Engineering |
| Datum | March 1992 |
| Jahrgang | 34 |
| Nummer | 3 |
| Seiten | 15-19 |
Literaturverz. |
no |
| Fußnoten | no |
| Fragmente | 2 |
| [1.] Svr/Fragment 071 12 - Diskussion Zuletzt bearbeitet: 2020-02-27 15:00:08 Schumann | Fragment, Gesichtet, KomplettPlagiat, SMWFragment, Schutzlevel sysop, Schuyler 1992, Svr |
|
|
|
| Untersuchte Arbeit: Seite: 71, Zeilen: 12-13, 17-32 |
Quelle: Schuyler 1992 Seite(n): 17, 18, Zeilen: 17: right col., 11 ff., 26 ff.; 18: left col., 3 ff. |
|---|---|
| 5.3.1 Decision Tree Analysis
A decision tree is a graphical extension of the Expected Value (EV) concept. [To calculate the expected value of a certain scenario, the values of the endpoints (possible scenario) are simply multiplied by the probabilities of reaching these endpoints. To consider the time value of money the value of each endpoint are discounted back to the present.] One of its strengths is that the tree diagram clarifies the decision problem with a visual approach. The diagram clearly shows all important decision elements, including contingencies (outcomes of chance events); decisions and alternatives, the logical sequence of decision points and chance events. One can often solve decision trees by hand, and the method is helpful at providing insights, especially related to the value of additional information and additional control. Branch and outcome values must be value measures (Present Value works, but, for example, internal rate of return (IRR) does not), and probabilities must be assessed or calculated for all chance event outcomes. One of the weaknesses of decision tree analysis is that one must represent all possibilities by a finite number of paths through the tree. This limits the practical number of random variables that can be accommodated, and one must use discrete distributions as approximations to continuous probability functions. Thus, range extremes are not represented. Other weaknesses of this approach include the fact that the analyst must limit the representation of uncertain time-spread events, such as for prices and inflation, to a few scenarios. Also, in [tree diagrams an output probability distribution (risk profile curve) is usually not obtained, only the EV.] |
[page 17]
1. Decision Tree Analysis—A decision tree is a graphical extension of the EV concept. Working from right to left, EV is back-calculated by:
The tree diagram clarifies the decision problem with a visual approach, one of its strengths. The diagram clearly shows all important decision elements, including contingencies (outcomes of chance events); decisions and alternatives; and the logical sequence of decision points and chance events. One can often solve decision trees by hand, and the method is helpful at providing insights, especially related to the value of additional information and additional control. One of the weaknesses of decision tree analysis is that one must represent all possibilities by a finite number of paths through the tree. This limits the practical number of random variables (chance events) that can be accommodated, and one must use discrete distributions as approximations to continuous probability functions; thus, range extremes are not represented. Other weaknesses of this approach include the fact that the analyst must limit the representation of uncertain time-spread events, such as for prices and inflation, to a few scenarios. Also, in tree diagrams an output probability distribution (risk profile curve) is usually not obtained, only the EV. [page 18] Branch and outcome values must be value measures [PV works, but, for example, internal rate of return (IRR) does not], and probabilities must be assessed or calculated for all chance event outcomes. |
The source is not given. |
|
| [2.] Svr/Fragment 072 01 - Diskussion Zuletzt bearbeitet: 2020-02-27 14:52:51 Schumann | Fragment, Gesichtet, KomplettPlagiat, SMWFragment, Schutzlevel sysop, Schuyler 1992, Svr |
|
|
|
| Untersuchte Arbeit: Seite: 72, Zeilen: 1-3 |
Quelle: Schuyler 1992 Seite(n): 17, 18, Zeilen: 17: right col., last paragraph; 18: left col., 1 ff. |
|---|---|
| [Also, in] tree diagrams an output probability distribution (risk profile curve) is usually not obtained, only the EV. However, with a modest extra effort, the distribution curve can be calculated by "rolling-forward" the tree. | [page 17]
Also, in tree diagrams an output probability distribution (risk profile curve) is usually not obtained, only the EV. However, [page 18] with a modest extra effort, the distribution curve can be calculated by “rolling-forward” the tree. |
The source is not given. |
|