Examination of decision support systems for composite CBA & MCDA assessments of transport infrastructure projects

This paper examines decision support systems (DSS) for composite appraisals of transport infrastructure projects comprising both cost-benefit analysis (CBA) and multi-criteria analysis (MCA). Two DSS are in this context examined and compared using a case study dealing with alternatives for a new high-speed railway line between the two cities, Norrköping and Bäckeby, in Sweden. Both systems are based on additive value functions and makes use of pair wise comparisons. The first system examined, which is widely used and based on acknowledged methods, comprises the REMBRANDT technique using pair wise comparisons for rating of the alternatives and swing weights for the determination of criteria weights. The results of the CBA are in this system converted into value function scores, assigned with a weight and included as an additional criterion in the MCA appraisal. Hence, the result of the system is a score for each alternative reflecting its performance in the composite appraisal. The second system examined, the so-called COSIMA approach, provides a framework for adding value functions determined in a MCA to impacts monetarily assessed in a CBA. The system makes use of the same methods as the latter for the rating of alternatives and the determination of criteria weights, namely the REMBRANDT technique combined with swing weights. However, this system does not convert the results of the CBA into value function scores. Instead the value functions computed in the MCA are added to the CBA results using a trade-off indicator. Hence, the result is a total rate for each alternative reflecting its attractiveness in the appraisal as a function of the trade-off between the CBA and MCA and their individual results. Thus, a sensitivity measure for the MCA’s influence on the final result of the composite analysis is achieved. The input for the two DSS examined was generated using the previous mentioned case study. A decision conference was set up where various stakeholders and decision makers under the guidance of a facilitator were producing input in form of their preferences. The purpose of the examination and comparison of the two systems is to determine which is the most appropriate for conducting composite appraisals of transport infrastructure projects. The first system provides a conventional widely used and theoretical well founded framework. The COSIMA framework is founded on a somewhat different set of axioms. However, the two systems provide the decision makers with the same kind of result, only expressed differently. The question treated in this context is hence if the COSIMA system provides the decision makers with some information that the conventional system does not provide and vice versa. The paper also treats issues regarding the use of scales within additive value functions. E.g. using a local scale defined by the alternatives that are under consideration set against a global scale defined by reference to a wider set of possibilities. Moreover, if an intrinsically important criterion in the appraisal does not differentiate much between the alternatives – if the minimum and maximum points on the value scale correspond to similar levels of performance – then that criterion should maybe be ranked quite low or even omitted from the appraisal. This issue particularly arises if the degree of differentiation differs significantly across criteria; on one criterion the reference points may represent low discrimination ability between the alternatives under consideration, whereas on another criterion the reference points may lead to high discrimination as concerns influence on the assessment score. It is examined if some lower boundary rule can be defined in this respect indicating when a criterion should be assigned with a low weight or even omitted from the appraisal. Finally, conclusions are drawn including a recommendation based on the case study for the most appropriate system for conducting composite appraisals of transport infrastructure projects, and research questions defining future work in the context of composite DSS and their use in decision making processes are set out.