Before you take a decision you can normally choose amongmany different alternatives. The alternatives can be rankeddepending on what net benefits they produce. The alternativewith the best net benefit is chosen. Within economicscost-benefit analysis may be used to decide which alternativeis likely to offer the best net benefit.
Some time after the decision it may be found that theselected alternative turned out not to be as favourable asexpected. The reason is often that there are uncertainties inthe decision process. Sometimes this is not a major problemsince it is possible to change the decision or the costconsequences may be small.
When the National Traffic Administrations make investmentslarge volumes of capital are involved. It is often difficult orvery expensive to change an investment when started. It istherefore essential to get an early knowledge about potentialuncertainties, which can affect the decision. Uncertainties canaccrue from incorrect input data. A reason for spreading ofuncertainties is the presence of dependencies among thevariables. The uncertainties are also usually dependent upontime.
This thesis describes some methods to take uncertaintiesinto account in cost-benefit analysis before a decision istaken. The aim is both to identify the uncertainties and tooutline methods for coping with uncertainties.
The methods rely on results in reliability theory and riskanalyses among others. Some statistical methods and discretedata models are also applied. The aim is to increase theprobability of selecting an alternative with high and stablenet benefits over a long time period.
The cost-benefit estimates are divided into smaller parts tomake it possible to locate uncertainties. Descriptive andanalytical models are recommended for handling the cost-benefitstructure and the uncertainties found in its components.Dependencies among variables and, in particular, timedependencies are analysed.
Throughout the thesis the various methods are applied to anexample referring to a road investment.
Stockholm: Infrastruktur , 2004. , 175 p.