Fusing non-conservative kinetic market models and evolutionary computing
Tipo
Artigo
Data de publicação
2020
Periódico
Physica A: Statistical Mechanics and its Applications
Citações (Scopus)
2
Autores
Luquini E.
Montagna G.
Omar N.
Montagna G.
Omar N.
Orientador
Título da Revista
ISSN da Revista
Título de Volume
Membros da banca
Programa
Resumo
© 2019This research establishes an identity between kinetic market models of econophysics and evolutionary algorithms of computer science. The fusion between the two approaches motivated a new market model with two basic operations: sampling and selection of states. The result is a non-conservative market that depends on the size of the sample set and the approach used to approximate the principle of energy conservation. This market exhibits complex dynamics with random walks for the sum of all the agents’ money and a scaling behavior for the money distribution in the population. Moreover, the fusion demonstrates how to add an evolutionary context to the kinetic market models and suggests a quasi-equilibrium version of those models. As a by-product, the work reveals a practical application as a new replacement rule for family competition evolutionary algorithms, which outperforms traditional ones in challenging combinatorial optimization problems.
Descrição
Palavras-chave
Assuntos Scopus
Combinatorial optimization problems , Evolutionary computing , Market model , Money distribution , Population diversity , Principle of energy conservation , Quasi equilibrium , Random Walk