A numerical study for optimal portfolio regime-switching model. I: 2D Black-Scholes equation with an exponential non-linear term. (Q507925)

From MaRDI portal
scientific article
Language Label Description Also known as
English
A numerical study for optimal portfolio regime-switching model. I: 2D Black-Scholes equation with an exponential non-linear term.
scientific article

    Statements

    A numerical study for optimal portfolio regime-switching model. I: 2D Black-Scholes equation with an exponential non-linear term. (English)
    0 references
    0 references
    0 references
    9 February 2017
    0 references
    A system of weakly coupled semilinear parabolic equations of optimal portfolio in a regime-switching model is studied. The model 2D Black-Scholes semilinear equation is discussed in Section 2. The following boundary value problem with Dirichlet boundary conditions \[ \begin{gathered} \frac{\partial u}{\partial\tau}- (1-\frac{a}2)\frac{\partial u}{\partial x_1}- (1-\frac{c}2)\frac{\partial u}{\partial x_2}- \frac12(a\frac{\partial^2u}{\partial x_1^2}+2b\frac{\partial^2u}{\partial x_1\partial x_2}+c\frac{\partial^2u}{\partial x_1^2})+ru\\ =\mathcal R\equiv -\frac{e^{-r\tau}}\alpha[e^{\alpha\phi u}-e^{-\alpha\phi u}+Z],\quad (\tau, x_1, x_2)\in Q_T,\tag{1}\\ u(0,x_1, x_2) = u^0(e^{x_1}, e^{x_2}),\quad (x_1, x_2)\in D;\\ u(\tau, L_W, x_2) = u_W(\tau, e^{x_2}), u(\tau, L_E, x_2) = u_E(\tau, e^{x_2}), (\tau, e^{x_2})\in[0, T]\times[L_S, L_N];\\ u(\tau, x_1, L_S) = u_S(\tau, e^{x_1}), u(\tau, x_1, L_N) = u_N(\tau, e^{x_1}), (\tau, e^{x_1})\in[0, T]\times[L_W, L_E] \end{gathered} \] is considered. Theorem 1 states: Suppose that the function \(u\in C(\overline{Q}_T)\cap C^{3,1}(\overline{Q}_T)\) satisfies in \(Q_T\) the problem (1) and the functions \(u^0\), \(u_W\), \(u_E\), \(u_N\), \(u_S\) are non-negative (non-positive). If the function \(Z\) is non-positive (non-negative) and the functions \(u^0\), \(Z\) are not both identically zero, then \(u\) is non-negative (non-positive). In Section 3 the numerical method for the problem (1) is described. In Section 4 conditions, which guarantee the sign preserving property (Theorem 2) and the convergence of the discrete problem (Theorem 6) are investigated. In Section 5 numerical experiments, which validate the theoretical results are presented. Conclusions and a plan for the future work are given in Section 6.
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    optimal portfolio regime-switching model
    0 references
    semilinear parabolic equation
    0 references
    sign preserving
    0 references
    monotonicity
    0 references
    stability
    0 references
    convergence
    0 references
    0 references
    0 references