1. In the presence of the delta-property, value additivity enables the notions of dominance, EVPI and EVSI to be extended to CE-dominance, CEVPI and CEVSI.
2. The CE Value of Perfect and Sample Information may exceed EVPI and EVSI respectively, in some cases by a wide margin.
3. A number of commonly used probability distributions, including the normal, exponential, gamma, uniform and histogram distributions, admit of closed form CE functions in terms of the distribution parameters.
4. A number of commonly used probability distributions, including the lognormal, beta and Weibull distributions, admit of algorithmic CE functions in terms of the distribution parameters. Some of these have been coded and tested in Visual Basic for Applications.
5. It is reasonable to expect that using the standard integration techniques (series expansions, change of variables, and integration by parts) together with standard numerical integration techniques (based on summing trapezodial areas), C, VB and VBA Certain Equivalent functions for ALL distributions currently offered in simulation packages will be developed in the near future.
6. Use of CE functions for asymmetric probability distributions may yield performance gains over those obtained for the normal assumption.
7. For commercial availability of these routines, check
my Web site at