We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
Applications of simulated annealing in market research.
- Authors
Sharot, Trevor
- Abstract
Simulated annealing (SA) is a statistical optimisation method with a wide variety of applications. The best known of these is the travelling salesman problem in operations research: a salesman has to visit a number of clients in different towns in a single trip - in which order should he make the calls to minimise the total distance travelled? With more than a small number of towns, the number of permutations is too large to permit an exhaustive search, so some method is required to find the optimum without excessive computation. Simulated annealing provides a practical solution to this and many other combinatorial problems. It works by starting with any non-optimal arrangement and making a series of incremental changes to arrive at the optimal, or a near-optimal, solution. In the case of the travelling salesman problem, each change would consist of exchanging two towns in the itinerary. If this reduces the total distance, then the exchange is confirmed; but it is also beneficial to accept some exchanges that increase total distance, since this permits a wider range of permutations to be explored. Despite its many successful applications, it seems that simulated annealing has not yet been applied within market research other than the developments reported here. This paper presents three applications of SA in market research. The first is in the design of matched subsamples, where sample elements are swapped between the subsamples to achieve optimum balance. The second is for choosing new members to recruit to panels, by swapping potential recruits on and off the list. The third is as a new method for performing cluster analysis, by swapping elements between clusters. In this application, simulated annealing provides insights into the properties of the popular k-means method.
- Subjects
MARKETING research; SIMULATED annealing; FAMOUS problems in probabilities; OPERATIONS research; TRAVELING salesman problem; CLUSTER analysis (Statistics)
- Publication
International Journal of Market Research, 2010, Vol 52, Issue 5, p577
- ISSN
1470-7853
- Publication type
Article
- DOI
10.2501/S1470785310201533