Fractional stochastic differential equations with applications to finance (Q713467)

Let \[ W_{t}^{H}:=\int_{0}^{t}(t-s)^{\alpha}dW_{s} \] be a Liouville fractional Brownian motion, where \(\alpha=H-1/2\), \(W_{s}\) is a standard Brownian motion. This paper deals with stochastic differential equations of the form \[ dX_{t}=b(t,X_{t})dt+\sigma(t)X_{t}dW_{t}^{H},\;X_0=x_0. \] It is assumed that the volatility \(\sigma\) is a deterministic bounded function on \([0,T]\) and the drift coefficient \(b\) is a measurable, in all its arguments, continuously differentiable function in \(x\) and \(| b(t,x)-b(t,y)|\leq L_0| x-y|\) for all \(x,y\in\mathbb{R}\), \(t\in [0,T]\), \(| b(t,x)|\leq L_0(1+| x|)\) for all \(x\in\mathbb{R}\) and for all \( t\in[0,T]\). The author proves that the considered fractional stochastic differential equation has a solution which is given by \[ X_{t}=Z_{t}/Y_{t}, \] where \(Y_{t}=\exp\left(-\int_{0}^{t}\sigma(s)dW_{s}^{H}\right)\) and \(Z_{t}\) is the unique solution of the ordinary differential equation \(dZ_{t}=Y_{t}b\left(t,Z_{t}/Y_{t}\right)dt\), \(Z_0=x_0\). Then, the author introduces a fractional semimartingale with two parameters \((H,\varepsilon)\in(0,1)\times(0,\infty)\) defined by \[ W_{t}^{H,\varepsilon}=\int_{0}^{t}(t-s+\varepsilon)^{\alpha}dW_{s} \] and considers the fractional semimartingale Black-Scholes model containing a bond \(dB_{t}=rB_{t}dt\), \(B_0=1\) and a stock \(dS_{t}^{\varepsilon}=\mu S_{t}^{\varepsilon}dt+\sigma S_{t}^{\varepsilon}dW_{t}^{H,\varepsilon}\), \(S_{0}^{\varepsilon}=S_0\). It is proved that this model has no arbitrage and is complete. The European call option price is obtained. Suppose that the price process follows the equation \(dS_{t}=S_{t}(\mu_{t}dt+\sigma dW_{t}^{H_1,\varepsilon_1})\), where \(\sigma\) and initial price are known positive constants, and drift coefficient follows the equation \(d\mu_{t}=\beta\mu_{t}dt+\nu dW_{t}^{H_2,\varepsilon_2}\), where \(\beta, \nu\) are unknown real constants, \((W_{t}^{H_1,\varepsilon_1},W_{t}^{H_2,\varepsilon_2})\) are independent. Let \(\pi_{t}\) be the number of shares invested in the stock at time \(t\), and let the wealth process be defined by \(dX_{t}^{x,\pi}=\pi_{t}dS_{t}\), where \(x\geq0\) is an initial wealth. The expected logarithmic utility maximization problem from terminal wealth for this model is solved.