 This paper proposes a novel hybrid inverse kinematics algorithm to solve the muscle-effort minimization problem in robotic applications. The algorithm uses a geometric Jacobian matrix to identify the null space of the geometric constraints, which allows for efficient exploration of the feasibility region. This is followed by a global search phase using an initial guess resampling strategy to find the optimal solution. The proposed method was tested on five right-handed subjects performing daily tasks, and the results showed that the robotic arm could accurately replicate the natural arm postures of the subjects. This article was authored by Kwam Linli, Yamsia, Shenzong Wang, and others.