A MULTI-OBJECTIVE ASSESSMENT OF INPUT–OUTPUT MATRIX UPDATING METHODS

This paper shows that important insights can be lost when assessing the relative performance of balancing methods solely based on individual optima. This is demonstrated through a multi-objective assessment. A trade-off curve between RAS and sign-preserving absolute differences (SPAD) is obtained based on the 60×60 Norwegian 2001 input–output table. The trade-off curve takes on a form that is close to a step function. This demonstrates that the solution surface around the RAS and SPAD optimums are very flat. Solutions can be identified that improve on the other objective or measure with little or marginal cost to the original objective function. Motivation for the assessment is provided, the technique applied is presented and the implications of the findings are discussed in an input–output and industrial ecology context.