Welcome to GCP-HOLO’s documentation!
Important
This is documentation is still a work in progress. Please reachout with any problems or missing information.
The GCP-HOLO documentation describes a method for 1DOF path synthesis of high order linkage graphs. GCP-HOLO stands for Graph Convolution Policy for High Order Linkage Optimization. We define high order linkage graphs as more than 3 active loops, although GCP-HOLO can generate linkages with less number of loops, other methods will perform better.
Path synthesis is a challenging task due to the non-linearities, combinatorial nature, and strict geometric constraints involved. GCP-HOLO iteratively grows a base linkage using Assur group 0DOF linkages. We are able to evaluate the new linkages quickly through symbolic kinematics described by Bacher et al. If you just niavely place a new joint and rigid bars to connect on to an existing 1DOF linkage, it is likely to be kinematically infeasible. To address this problem we introduce scaffold nodes which are potential initial joint locations. This gives the problem a large, but fixed action space to grow the linkage. The project allows you to train a Deep Reinforcement Learning (DRL) agent to learn a design heuristic for the path synthesis problem. This project supports 3 types of agents from the Stable Baselines3 library: PPO, A2C, and DQN, as well as a random search option as well.