No label defined (Q4356587)

We consider a model of \(n\) risky and one risk-free asset in which the dynamics of the prices of the risky assets are governed by logarithmic Brownian motions (LBM). One of the avenues of research deals with the transaction cost problem: Consider an investor who has an initial wealth invested in Stocks and Bonds and who has the ability to transfer funds between the assets. This transfer involves brokerage fees (transaction costs). In one type of model the proceeds from these operations are used to finance the consumption and the objective is to maximize the cumulative expected utility of consumption over a finite or infinite time horizon. Sometimes, a similar problem is considered but on a finite horizon and without consumption, or the objective is to maximize the long run average growth of wealth. Mathematically, these problems can be formulated as singular stochastic problems. Dynamic programming methodology leads to variational inequality (or dynamic programming equation) for the value function of elliptic, parabolic or ergodic type, that is a nonlinear partial differential equation of second-order. Solution of these equations provides the optimal investment strategy in feedback form which means that the optimal control is a function of the current state of the system. In general, the value function is not smooth enough to satisfy the dynamic programming equation in the strong sense. Therefore a weak formulation of solutions is necessary: the concept of viscosity is a powerful tool which is well adapted to these equations. Moreover, except for special cases such as the Merton problem (no transaction costs), dynamic programming equations cannot be solved explicitly and we thus have to turn to numerical solutions. The aim of this article is to present some of the tools used for dynamic optimization of portfolios. The article is organized as follows: the mathematical formulation of the problem in terms of the stochastic control problem is given in Section 2. Two cases are considered: maximization of a function of the net wealth on a finite horizon and maximization of the long run average growth of wealth. We state in Section 3 the variational inequalities satisfied by the value function in both parabolic and ergodic cases. In Section 4, numerical methods are presented to solve the dynamic programming equations: that is, (i) restriction to a bounded domain; (ii) discretization; (iii) solution of the discretized problem. Finally, in the last section, numerical results are given in the case of one bank account and two risky assets: optimal transaction policies are obtained and risk-return characteristics are analyzed. For the entire collection see [Zbl 0867.00036].