The consumer should be able to buy only one good, choosing between the one provided by the various companies. The parameters that I want to consider are: the price, the company's visibility in the market, the quality of the product and the quality perceived by the consumer.
You might find some ideas here: Saemundsson, R., K. R. Thórisson, G. R. Jonsdottir, M. Arinbjarnar, H. Finnsson, H. Gudnason, V. Hafsteinsson, G. Hannesson, J. Ísleifsdóttir, Á. Th. Jóhannsson, G. Kristjánsson & S. Sigmundarson (2006). Modular Simulation of Knowledge Development in Industry: A Multi-Level Framework. Proc. of WEHIA – 1st International Conference on Economic Sciences with Heterogeneous Interacting Agents, 15-17 June, University of Bologna, Italy.
LINK: http://xenia.media.mit.edu/%7Ekris/ftp/KnowledgeSimulation-WEHIA06.pdf
I could suggest you to take a look at the papers from Lamjed Ben Said, who has worked with France Telecom Marketing Department around 2002, building agent-based models of consumers facing a choice between three brands in a small (virtual) city. Two references among others :
Lamjed Ben Said, Thierry Bouron: Multi-Agent Simulation of Virtual Consumer Populations in a Competitive Market. SCAI 2001: 31-43
Lamjed Ben Said, Thierry Bouron, Alexis Drogoul: Agent-based interaction analysis of consumer behavior. AAMAS 2002: 184-190
Thank you all for the contributions indicated!
My initial idea was the following ...
For each customer, or customer type, I would like to define a function utility u (p0, p1, ..., pn) that takes as arguments the price p0 and the other n parameters p1, ..., pn that describe other characteristics of the offer. The customer then chooses the offer that maximizes the utility function.
Now I would like a suggestion on the "form" should have the function, whereas it should be consistent with different types of customers.
The most obvious way would be to consider a linear utility function
u (p0, p1, ..., pn) = a0 + a1 p0 p1 + .. + an pn
where the terms in the weights that are attributed to each parameter.
These weights may also be negative, for example for the price and should also vary according to the type of customer.
Of course, you could assign an absolute weight more parameters to more 'important.
In the extreme case for example in which a customer is interested only at the price, I would place a0 = -1 and all other parameters are equal to zero.
Do you have any reference to suggest?
See http://www.economicswebinstitute.org/essays/dyncomp.htm If you like, I can send you the full code of the model.
My suggestion is to use the Cox Regressions (survival analyzes) because the covariates in your model are time-dependent and a simple OLS model can not deal with time-dependent variables and censoring. This kind of analysis’s can be done by using the statistical program SPSS. Tutorials for the Cox Regressions in SPSS can be easily found on web.
This may help "Comparing Cartel Behavior: A Simulation Analysis with
the System of Cartel Markers (SCM)"
- Badges
- Science topic
- Similar topics
- Geoscience
- Asia
- India