Mathematics > Probability
[Submitted on 23 Oct 2022 (v1), last revised 19 Jun 2023 (this version, v4)]
Title:Large Deviations Theory of Increasing Returns
View PDFAbstract:An influential theory of increasing returns has been proposed by the economist W. B. Arthur in the '80s to explain the lock-in phenomenon between two competing commercial products. In the most simplified situation there are two competing products that gain customers according to a majority mechanism: each new customer arrives and asks which product they bought to a certain odd number of previous customers, and then buy the most shared product within this sample. It is known that one of these two companies reaches monopoly almost surely in the limit of infinite customers. Here we consider a generalization [G. Dosi, Y. Ermoliev, Y. Kaniovsky, J. Math. Econom. 23, 1-19 (1994)] where the new customer follows the indication of the sample with some probability, and buy the other product otherwise. Other than economy, this model can be reduced to the urn of Hill, Lane and Sudderth, and includes several models of physical interest as special cases, like the Elephant Random Walk, the Friedman's urn and other generalized urn models. We provide a large deviation analysis of this model at the sample-path level, and give a formula that allows to find the most likely trajectories followed by the market share variable. Interestingly, in the parameter range where the lock-in phase is expected, we observe a whole region of convergence where the entropy cost is sub-linear. We also find a non-linear differential equation for the cumulant generating function of the market share variable, that can be studied with a suitable perturbations theory.
Submission history
From: Simone Franchini Dr. [view email][v1] Sun, 23 Oct 2022 01:19:34 UTC (112 KB)
[v2] Sun, 2 Apr 2023 18:12:17 UTC (692 KB)
[v3] Sat, 20 May 2023 10:12:55 UTC (693 KB)
[v4] Mon, 19 Jun 2023 18:00:04 UTC (696 KB)
Current browse context:
math.PR
Change to browse by:
References & Citations
Bibliographic and Citation Tools
Bibliographic Explorer (What is the Explorer?)
Connected Papers (What is Connected Papers?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)
Code, Data and Media Associated with this Article
alphaXiv (What is alphaXiv?)
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Hugging Face (What is Huggingface?)
Papers with Code (What is Papers with Code?)
ScienceCast (What is ScienceCast?)
Demos
Recommenders and Search Tools
Influence Flower (What are Influence Flowers?)
CORE Recommender (What is CORE?)
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.