Dissertation
Optimal execution problem: a mean field game and fictitious play study
Fecha
2020-06-26Autor
Birman, Bernardo
Institución
Resumen
In this work we present the main ideas of the optimal execution problem and a couple of different approaches for its modelling. We focus on the Mean-Field Game approach, and observe how restrictive it can be as we only can obtain a explicit solution for the simplest case. We model a learning algorithm through the concept of the fictitious play, and check that our strategy generated by the model outperforms some different, naive, strategies. Neste trabalho apresentamos as ideia principais do problema de execução ótima e algumas abordagens para sua modelagem. Nós focamos na modelagem de Mean-Field Game e observamos quão restritiva ela pode ser, por só obtermos uma solução explícita para o caso mais simples. Nós modelamos um algoritmo de aprendizagem através do conceito de fictitious play e checamos que a estratégia gerada pelo nosso modelo apresenta uma performance superior a estratégias mais ingênuas.
Ítems relacionados
Mostrando ítems relacionados por Título, autor o materia.
-
Real Options Approach to optimizing long term mine planning
Epstein Numhauser, Rafael David; UNIVERSIDAD DE CHILE (2003)This thesis uses a dynamic programming approach to sequence the extraction of different blocks of a mine when there is uncertainty on the spot price of the output of the mining activity. To select the parameters of such ... -
Real Options Approach to optimizing long term mine planning
Saure Valenzuela, Denis Roland -
NECESSARY AND SUFFICIENT OPTIMALITY CONDITIONS FOR CONTINOUS-TIME MULTIOBJECTIVE OPTIMIZATION PROBLEMS
Batista Dos Santos, Lucelina; Brandao, A J; De Oliveira, Valeriano Antunes; Rojas Medar, Marko Antonio