A method is proposed which accomplishes a whole task consisting of plural subtasks by coordinating multiple behaviors acquired by a vision-based reinforcement learning. First, individual behaviors which achieve the corresponding subtasks are independently acquired by Q-learning, a widely used reinforcement learning method. Each learned behavior can be represented by an action-value function in ...

Optimization problems in chemical engineering often involve complex systems of nonlinear DAE as the model equations. The direct multiple shooting method has been known for a while as a fast off-line method for optimization problems in ODE and later in DAE. Some factors crucial for its fast performance are briefly reviewed. The direct multiple shooting approach has been successfully adapted to t...

This paper describes an adaptive algorithm for optimal control of time-dependent partial di erentialalgebraic equation (PDAE) systems. A direct method based on a modi ed multiple shooting type technique and sequential quadratic programming (SQP) is used for solving the optimal control problem, while an adaptive mesh re nement (AMR) algorithm is employed to dynamically adapt the spatial integrat...

We present here an alternative method for calculating magnifications in gravitational lensing calculations – the Ray Bundle method. We provide a detailed comparison between the distribution of magnifications obtained compared with analytic results and conventional ray-shooting methods. The Ray Bundle method provides high accuracy in the weak lensing limit, and is computationally much faster tha...

We describe a robust shooting algorithm to model backward-pumped Raman amplifiers. This algorithm uses a continuation method and a Jacobi weight in conjunction with the shooting algorithm. We compare this algorithm to the commonly used relaxation algorithm. We find that the shooting algorithm is more flexible, in that it can be applied to amplifiers in which one fixes the gain, in contrast to t...

Minimum time and minimum energy point-to-point trajectories for an industrial robot of the type Manutec r3 are computed subject to state constraints on the angular velocities. The numerical solutions of these optimal control problems are obtained in an eecient way by a combination of a direct collocation and an indirect multiple shooting method. This combination links the beneets of both approa...

Multiple shooting methods for solving optimal control problems governed by ODEs have been extensively studied in past decades. However, their application large-scale PDE-based still faces many challenges, including the difficulty of large scale equality constrained optimization an efficient parallelizable way. The current work proposes and analyzes a new parallel-in-time multiple algorithm para...