DeepXube Basics

Overview

The objective of DeepXube is to automate the solution of pathfinding problems by using deep reinforcement learning to learn heuristic functions that guide heuristic search to solve problems. As a result, user implementation is reduced to the definition of the domain, neural network input, and heuristic function. Furthermore, since DeepXube comes with common neural networks, such as residual neural networks, the implementation of a heuristic function architecture may not be necessary.

Tip

See Creating a Custom Domain and Heuristic Function for implementing custom domain, neural network input, and heuristic function architecture classes.

Domains

Training

Overview of the DeepXube training pipeline.