Hybrid evolutionary algorithms for detecting parameter stability in economic systems

The paper is primarily intended to be a study of the initial values' influence on finding the optimal solution for an economic macromodel. Empirical trials showed a high sensitivity of the model's output to small variations of the input values required to run the model. A second goal of our approach is rising some questions on the adequacy of different objective functions for the same economic model. Third, we tackle the problem of instability of coefficients involved in econometric equations. We propose finding several bootstrap confidence intervals and then applying hybrid evolutionary algorithms for estimating these coefficients instead of usual 'ordinary least squares', 'two-stage least squares', or other classical approaches.