Existence and uniqueness of solutions to a quasilinear parabolic equation with quadratic gradients in financial markets (Q2486634)

The authors study the first initial boundary-value problem for a quasilinear parabolic equation which models an optimal portfolio in an incomplete financial market consisting of risky assets and and non-tradable state variables. Consider the market consisting of \(d\) risky assets \(S_i\) and \(d'\) non-tradable state variables \(S'_k.\) To construct an optimal value function which allows to compute the optimal trading strategy they consider the equation \[ \partial_tu-{1\over 2}\sum_{i,j=1}^n \widetilde c_{ij}\partial_i\partial_j u= \sum_{i,j=1}^n\widetilde \mu_i\partial_iu-q\sum_{i=1}^d(\mu_i-rS_i)\partial_i u -{q\over 2}\beta(u)^2+pr +\sum_{i,j=1}^n\widehat c_{ij}(u)\partial_iu\partial_ju. \tag{1} \] Here \(n=d+d'\), \(\widetilde C=(C, C'), C=(c_{ij}(S,S',t,u)\), \(i,j=1,\dots,d)\) is the diffusion matrix of risky assets, \(C'=(c'_{kl}(S,S',t,u)\), \(k,l=1,\dots,d')\) is the diffusion matrix of the non-tradable state variables, \(\widehat C=({1\over 2(p-1)}C, {1\over 2}C'),\) \(\widetilde\mu=(\mu, \mu')\) while \( \mu(S,S',t), \mu'(S,S',t)\) and \(r(S,S',t)\) are respectively the expected returns and the riskless interest rate. Finally, \(\beta=(\mu-rS)^TC^{-1}(\mu-rS)\) is the square of risk premium and numbers \(p, q\) satisfy the relation \({1\over p}+{1\over q}=1\). In the domain \(\Omega\times\Omega'\times(0,T)\subset \mathbb R^n\times (0,T)\) for the equation (1) the authors consider both the Cauchy problem with initial data \[ u(S,S',0)=u_0(S,S') \tag{2} \] ( for \(\widehat\Omega=\Omega\times\Omega'=R^n\) ) and the initial boundary-value problem with the boundary condition \[ u(S,S',t)=u_D(S,S',t), \quad (S,S',t)\in \partial \widehat\Omega\times (0,T).\tag{3} \] Under some additional assumptions of the data of the above problems the authors apply a fixed point argument to prove the existence of a solution \(u^\varepsilon\) of a linearized problem corresponding to (1), and check that it is uniformly bounded in the Sobolev norm and converges to the solution of the original problem. The uniqueness of the solution to (1)--(3) is also proved. In the last section the authors give an application and a numerical example.