Hybrid Differential Dynamic Programming, HDDP

Author: Tianjiang Shuo

Website: https://cislunarspace.cn

Definition

A trajectory optimization algorithm that extends differential dynamic programming (DDP) with augmented Lagrangian constraint handling and trust-region regularization. Proposed by Lantoine and Russell, HDDP adjoins terminal constraints using both Lagrange multipliers and quadratic penalty terms, converting constrained optimization into a sequence of unconstrained subproblems solved iteratively. It handles equality and inequality constraints with greater tolerance for poor initial guesses than standard DDP. As a local optimizer, it does not explore multiple local minima.

Application Value

The 混合差分动态规划 concept provides technical support and analytical methods for cislunar space research and lunar exploration missions.

References

  • Aziz et al. 2019, JGCD, DOI: 10.2514/1.G003617
Improve this page
Last Updated:
Contributors: ouyangjiahong, Ou Yang Jiahong