Improved genetic algorithm for estimation of parameters of demand functions

The method of econometrics for estimation of the parameters of demand functions is the probabilistic and statistic method for least square estimation. It is sometimes possible that the method of econometrics does not obtain the optimal parameters. To solve the problem, this paper puts forward improved genetic algorithm (IGA) which is suitable for the estimation of the parameters of nonlinear and linear demand functions. IGA has the mechanism combining the global optimization with the local optimization, and IGA is capable of revising the parameters many times under the instruction of system errors. So far as the linear estimation of parameters is concerned, the time complexity of IGA is O(n). It is also less than that of the method of econometrics which is O(n/sup 2/). The simulation example shows that the system error of IGA is 0.405680 which is less than 0.41057 derived from the method of econometrics. Therefore IGA is efficient.