使用粒子群算法解决期权定价模型参数校准问题——以heston模型为例

期权定价模型的参数校准问题是一个常见的难题，以heston 模型为例，定价时需要估计6 个参数，参数估计问题实质上是高维非线性规划问题，由于估参函数的性质不好，一般的估参方法常常失效。使用粒子群（PSO）智能算法可以改善该模型的参数校准问题，因为粒子群算法具有内在随机性，因此参数估计中的局部极小值问题可以被较好地解决。使用2017 年12 月20 日的香港恒生指数期权作为估计样本，并对2017 年12 月25 日的期权进行样本外预测，数值结果表明使用heston 模型对期权进行定价并配合粒子群算法估计参数具有良好的定价效果。

Solving The Problem of Parameter Estimation in Heston Model with Particle Swarm Algorithm

The parameter calibration of option pricing model is a common problem. Taking Heston model as an example, six parameters need to be estimated when pricing. The problem of estimation is essentially a high-dimensional nonlinear programming problem. Due to the bad nature of estimated parameter function, general estimating methods often fail.Using particle swarm intelligence algorithm (PSO) to improve the parameter estimation outcome of heston option pricing model. Because of the inherent randomness in particle swarm algorithm, local minimum value problem in parameter estimation can be solved. Using HIS options data on December 20, 2017 as the sample of estimation and HIS options on December 25, 2017 as the sample of prediction. The numerical results show that using heston model to price the options combing the particle swarm optimization algorithm to estimate parameters will have a good pricing outcome.