Cash balance optimization with genetic algorithms: a study of cross and mutation relationship in Miller-Orr model

Abstract

This work has the objective to apply genetic algorithms in cash balance policy definition. This financial problem was initially treated by Baumol (1952) and Tobin (1956) which work applies deterministic models to inventory control in enterprises cash balance. Further, Miller and Orr (1966) enhanced the problem approach introducing a stochastic model which no longer defines the optimal cash balance, but bands of oscillation. This work proposes an evolutionary model methodology with genetic algorithms, in different approaches to cash balance optimization, using the premises showed in literature. For this, simulations are used in model support and validation. The results shows that genetic algorithms can be very useful in Miller-Orr model parameterization, with good results in this field of problem, and the optimal solution has a strong association with mutation evolutionary process. This paper let future perspectives of better application of genetic algorithms in cash balance optimization problem.