rushingfox / evolutionary-algorithm-to-solve-the-inverse-problem-of-differential-equations

Wuhan University Bachelor Dissertation: the Application of Evolutionary Algorithms in the Parameter Optimization of Differential Equations

Evolutionary-Algorithm-to-solve-the-Inverse-Problem-of-Differential-Equations

This repository is the code of Jianhao Wu's Bachelor Dissertation in the School of Computer Science, Wuhan University: The Application of Evolutionary Algorithms in the Parameter Optimization of Differential Equations.

This thesis focuses on the application of evolutionary algorithms to the inverse problem of continuous function parameters for elliptic and parabolic differential equations.

Not only reproduced the old algorithm but also proposed a stepwise evolutionary algorithm based on this algorithm, i.e., the original algorithm is still used for a wide range of search activities at the early stage of evolution, and an elite variational operator designed by me is added for further parameter optimization at the later stage of evolution, which effectively improves the upper limit of the accuracy of the original algorithm and also reduces the possibility of falling into a local optimum.