A Study Of Real Option Pricing Method Based On ANN

Options theories provided a new tool for investment decision-making. In recent years, many overseas and domestic experts had paid attention to research options pricing. But these methods couldn't get rid of not only the subjectivity and fuzzy qualities existing in pricing process but also the restraint of Black-Scholes model on the choice of the variable, and the accuracy of assessed results had not quite improved. It will be very helpful to investors, if we could set up an objective scientific real option pricing method.The paper planed to probe into a new pricing method, which was not only designed to resolve the defaults of present pricing methods, but also linked to the practice of investment project. This paper researched on real option pricing through combining BP artificial neural network and GA.This paper's study was emphasized on:1. Set up BP neural networkSet up BP network by MATLAB toolbox. Chosen transfer function among every layer, training error and learning rate were included.2. GA optimized BP neural networkFirst, selected function for evaluating. Second, used MATLAB toolbox to design GA (chosen selection methods, crossover type, and mutation probability).GA could get rid of redundant node and branch effectively from BP network, and optimized it.3. Tested the pricing modelThrough analyzing real option existing in the investment project in detail to define the variables for network input. If there had much difference between simulation result and the value of investment, we should set and optimize BP network again. This course would not stop until we got the simulation result which was satisfied.This paper used MATLAB toolbox to edit programs, which could realize all kinds operations of ANN and GA. After training samples with the programs, we could set up real option pricing model based on ANN. Furthermore, we tested this model with an example.Last, compared the results which were calculated by Black-Scholes model and real option pricing model with the investment value, we could find that the result calculated by BP network was nearly the same as it. The simulation result has verified the validity of this method, and we could use this method to evaluate investment project.Set up real option pricing model based on ANN was the progeny of this paper. This paper's innovation was to use BP network and GA to price real option.We could draw conclusions as follows:1. BP model can quite improve the accuracy of pricing result;2. Need to confirm the number of the hidden layer of neuron rationally;3. Should confirm population size rationally while optimizing ANN;4. Need to design crossover operator rationally while optimizing ANN.