A Simulated Annealing Heuristic for a Bicriterion Partitioning Problem in Market Segmentation

Abstract

K-means clustering procedures are frequently used to identify homogeneous market segments on the basis of a set of descriptor variables. In practice, however, market research analysts often desire both homogeneous market segments and good explanation of an exogenous response variable. Unfortunately, the relationship between these two objective criteria can be antagonistic, and it is often difficult to find clustering solutions that yield adequate levels for both criteria. The authors present a simulated annealing heuristic for solving bicriterion partitioning problems related to these objectives. A large computational study and an empirical demonstration reveal the effectiveness of the methodology. The authors also discuss limitations and extensions of the method.

Scatter tabu search for multiobjective clustering problems

“…The severity of this problem increases markedly when three or more criteria are considered. This is the approach used by Brusco et al (2002) in their simulated annealing procedure. However, the approach has some severe limitations when dealing with combinatorial multiobjective optimization problems.…”

“…Given a set of customers, a set of descriptor variables and a response variable, the problem is to find a clustering of customers that maximizes both a measure of homogeneity and the explanation power of the response variable. As described in Brusco et al (2002), the bi-criterion problem consists of finding partitions that simultaneously maximize the proportion of variation explained by the descriptor variables and the response variable. The 'classical' market segmentation problem deals with creating homogenous groups and therefore deals with the single objective of maximizing the variation explained by the descriptor variables.…”

Amalgamation of partitions from multiple segmentation bases: A comparison of non-model-based and model-based methods

“…When different partitions of customers are obtained for different segmentation bases, perhaps with varying number of segments across partitions, a natural problem that arises is the establishment of a single consensus partition that best reflects an amalgamation of the individual partitions (Krieger and Green, 1999). 3 The potential for segmenting consumers on different bases has been identified by a number of authors (Krieger and Green, 1996;Ramaswamy et al, 1996;Brusco et al, 2002Brusco et al, , 2003Andrews and Currim, 2003a;Dolnicar and Grün, 2008;Tkaczynski et al, 2009). Unfortunately, segment partitions for different bases often do not correlate well with one another.…”

“…Unfortunately, segment partitions for different bases often do not correlate well with one another. For example, household responses to product and marketing mix characteristics often do not correlate well with household characteristics (Wind, 1978;Green and Krieger, 1991;Gupta and Chintagunta, 1994;Brusco et al, 2002), resulting in segmentation that is ineffective.…”

A Tabu Search Approach to Clustering

“…Methods that use and guide other heuristics in order to produce solutions beyond the local optima that are normally produced are meta-heuristics [3]. A popular meta-heuristic for MSSC Clustering problems is simulated annealing (SA); see [2]. It produces good cluster solutions and is relatively easy to implement.…”