- Abstract. Optimal pricing of European call option is described by linear stochastic differential equation with mean income and volatility. Trading strategy is given by a twin of stochastic variables for which self-financial condition is not postulated. Optimal pricing will be deformed to be adoptive to trading strategy employing martingale property where stochastic integral w.r.t. analytical solution of Black-Scholes type stochastic differential equation will be employed.

1 Stochastic Optimization

.

1.1 Black-Scholes Formula

Black-Scholes' formula starts from heat equation

the solution is given by

and corresponding stochastic differential equation is

where
${B}_{t}$
denotes Brownian motion. In case there is a drift term we have

whose solution is

where
$c$
is a velocity.

$$\begin{array}{c}\frac{\partial u}{\partial t}=\frac{1}{2}\frac{{\partial}^{2}{u}^{2}}{{\partial}^{2}{x}^{2}},\end{array}$$ | (1.1) |

$$\begin{array}{c}u=\frac{1}{\sqrt{2\pi t}}{e}^{-\frac{{u}^{2}}{2t}},\end{array}$$ | (1.2) |

$$\begin{array}{c}d{u}_{t}=d{B}_{t},\end{array}$$ | (1.3) |

$$\begin{array}{c}d{u}_{t}=-cdt+d{B}_{t},\end{array}$$ | (1.4) |

$$\begin{array}{c}u=\frac{1}{\sqrt{2\pi t}}{e}^{-\frac{(u-ct{)}^{2}}{2t}},\end{array}$$ | (1.5) |

Let us modify the last equation to be

where
$\alpha $
is volatility. This equation can be integrated to be

Let

and applying Ito formula we obtain

where let

which represent averaged income, and this can be integrated to be

This is the optimal pricing of European call option known as Black-Scholes formula.
${}^{1}$
It will be worth to note that eq.(1.6) include perturbed drift as a result there is an underlying displacement.

$$\begin{array}{c}{x}_{t}=cdt+\sqrt{\alpha}d{B}_{t},\end{array}$$ | (1.6) |

$$\begin{array}{c}{x}_{t}={\int}_{0}^{t}c(t-s)\sqrt{\alpha}d{B}_{s}.\end{array}$$ | (1.7) |

$$\begin{array}{c}{X}_{t}={X}_{0}{e}^{{x}_{t}},.\end{array}$$ | (1.8) |

$$\begin{array}{c}d{X}_{t}=\mu {X}_{t}dt+\sqrt{\alpha}{X}_{t}d{B}_{t},\end{array}$$ | (1.9) |

$$\begin{array}{c}\mu =c+\alpha ,\end{array}$$ | (1.10) |

$$\begin{array}{c}{X}_{t}={e}^{\sqrt{\alpha}{B}_{t}+\frac{(\mu -\alpha )}{2}t}.\end{array}$$ | (1.11) |

1.2 Trading Strategy

Trading strategy is a twin of stochastic variables denoted by
${}^{2}$

In this equation
${a}_{t}$
and
${b}_{t}$
are given by as defined in the space of square integral function
${L}^{2}$

and

where
$r$
denotes interest ratio and
${\beta}_{t}$
is pricing process of bond given by

Using these notation the portfolio is given by

$$\begin{array}{c}({a}_{t},{b}_{t}).\end{array}$$ | (1.12) |

$$\begin{array}{c}{a}_{t}d{X}_{t}=\mu {X}_{t}dt+\sqrt{\alpha}{X}_{t}d{B}_{t}\end{array}$$ | (1.13) |

$$\begin{array}{c}d{b}_{t}=r{b}_{t}{\beta}_{t}dt,\end{array}$$ | (1.14) |

$$\begin{array}{c}d{\beta}_{t}=r{b}_{t}{\beta}_{t}dt.\end{array}$$ | (1.15) |

$$\begin{array}{c}{V}_{t}={a}_{t}{X}_{t}+{b}_{t}{\beta}_{t}.\end{array}$$ | (1.16) |

1.3 Adaptive Optimization

Let us consider a stochastic integral of optimal pricing w.r.t. trading strategy as follows

If r.h.s. does not change pricing is called adoptive to trading strategy. This defines a martingale.
${}^{3}$
This means that pricing is globally optimized to trading strategy.

$$\begin{array}{c}{X}_{t}={\int}_{0}^{t}{X}_{s}d\{{a}_{s},{b}_{s}{\}}_{s}.\end{array}$$ | (1.17) |

This equation can be solved by iteration as follows.

repeat until

continue This provides an algorithm implementable into digital computer where the optimal pricing adoptive to trading strategy in which r.h.s denotes a conditional probabilty.

$$\begin{array}{c}\mathbf{E}\left({\mathbf{X}}_{\mathbf{m}}\right|\{{\mathbf{a}}_{\mathbf{n}},{\mathbf{b}}_{\mathbf{n}}{\}}_{\mathbf{n}})={\mathbf{X}}_{\mathbf{n}}<\epsilon ,\mathbf{a}.\mathbf{e}.\mathbf{m}>\mathbf{n}\end{array}$$ | (1.18) |

In this case, trading strategy needs not be self-financing.

2 Concluding remarks

Optimal pricing for risky asset adoptive to trading strategy was formulated in terms of Ito type stochastic differential equation. It should be noted, however, that there is a limitation in present formulation that it rely on Brownian motion.

As a result it depends on Gaussian process based on central limit theorem which states summation of independent variable is governed by Gaussian distribution.

The shortcomings arises from the fact that variables are not independent in the real world.

References
Department of Business Administration, Matsuyama University, 2-4 Bunkyo, Matsuyama, Ehime 790 8578, Japan E-mail address : tofukumi@cc.matsuyama-u.ac.jp

- T.Mikosh, Elementary Stochastic Calculus with Finance in View (World Scientific Pub. 1998).
- D.Duffie, Dynamic Asset Pricing Theory (Princeton University Press 1996).
- T.Hida, Brownian Motion (Springer-Verlag 1980).