Application of analytic hierarchy process to prioritise urban transport options – comparative analysis of group aggregation methods

Abstract

This paper presents a comparative analysis of different group aggregation methods adopted in analytic hierarchy process by testing them against social choice axioms with a case study of Delhi transport system. The group aggregation methods and their correctness were tested while prioritising the alternative options for energy efficient and less polluting transport system in Delhi. It was observed that among all group aggregation methods, geometric mean method, the most widely adopted group aggregation method of analytic hierarchy process, showed poor performance and failed to satisfy the most popular ‘Pareto optimality and non-dictatorship axiom’ raising questions on its validity as group aggregation method. All other group aggregation methods namely weighted arithmetic mean method with varying weights and equal weights and arithmetic mean of individual priorities resulted in concurring results with the individual member priorities. It was observed that weighted arithmetic mean method with equal weights aggregates individual priorities better compared to arithmetic mean method with varying weights. Arithmetic mean of the priorities by individual members of the group showed minimum deviation from the group consensus making it most suitable and simple method of aggregation.