On an Inner-Adaptive Multiple Shooting Method for Optimal Control Problems
摘要
Traditional multiple shooting method for optimal control problems often lacks flexibility in handling control switching and typically depends on numerical differentiation to compute gradients. This approach can be computationally inefficient, particularly under high-precision requirements. This paper presents an inner-adaptive multiple shooting method that incorporates internal switching points and a time-scaling transformation strategy to enhance adaptability. A gradient formula for the approximate problem is derived, supported by rigorous theoretical proof based on the variational principle. Numerical experiments demonstrate that the proposed method improves computational efficiency while offering greater flexibility in control strategies.