The Worth of Cardinal Information in MCDM – a Guide to Selecting Weight-Generating Functions

Danielson, M. & Ekenberg, L. ORCID: (2021). The Worth of Cardinal Information in MCDM – a Guide to Selecting Weight-Generating Functions. In: Innovation for Systems Information and Decision (INSID). Eds. de Almeida, A.T. & Costa Morais, D., pp. 20-35 Springer, Cham. ISBN 978-3-030-91768-5 10.1007/978-3-030-91768-5_2.

Full text not available from this repository.


There exist a variety of methods for extracting weights and values in multi-criteria decisions based on different rankings. However, it is difficult to determine which the useful ones are and how they correspond to the decision makers’ perceptions, if at all. How do we know that what we are using is really significant, especially in situations when the decision bases are vague? One category of methods that has proved relatively successful is to use so-called generated surrogate weights that are, in some sense, meant to represent rankings and there are various suggestions as to how best to distil them from input information. In this paper, using a number of simulations, we review some leading algorithms for automatic weight generation without external parameters besides cardinal and ordinal rankings and provide some guidelines for selecting a surrogate weight-generating function for MCDM applications, in ordinal as well as cardinal information settings. We also propose an alternative with some attractive properties compared with the existing ones.

Item Type: Book Section
Additional Information: Third Innovation for Systems Information and Decision Meeting, INSID 2021, Virtual Event, December 1–3, 2021, Proceedings
Uncontrolled Keywords: Multi-criteria decision analysis; Surrogate numbers; Robustness; Rank order
Research Programs: Advancing Systems Analysis (ASA)
Advancing Systems Analysis (ASA) > Cooperation and Transformative Governance (CAT)
Depositing User: Michaela Rossini
Date Deposited: 17 Dec 2021 09:48
Last Modified: 29 Apr 2024 12:37

Actions (login required)

View Item View Item