This repository contains my submission of Speedy and weighted A* algorithms to the GPPC in 2015

Found at movingai.com/GPPC, the competition is an exercise in various pathfinding approaches that are compared against each other using several metrics, including speed, memory consumption, and suboptimalility. This competition is run by Nathan Sturtevant of the University of Denver.

These algorithms were chosen to focus on a suboptimal approach to pathfinding. These approaches are much faster than optimal solutions as they reduce the need to exhaust all optimal paths along the way to their goal. As such, they have greatly improved performance at the cost of a bounded suboptimality.

Each algorithm (Speedy and multiple variations of weighted A*) are contained in their own directories. Each directory contains a makefile that will compile the algorithm implementation along with Entry.h (a header provided by the GPPC for API calls required for the competition). Each directory also contains main.cpp, another provided by the competition that will exectute the algorithm and display metrics when invoked with various inputs

The GPPC provides several examples of maps and scenarios to be tested. Maps are grids up to 2048x2048 in size, and scenarios are a list of start and goal state that the algorithm must plan for - along with some metrics to test the provided plan. These maps are generic pathfinding grids and mazes, but also contain maps from the game Warcraft 3, Starcraft 2, and Dragon Age Origins.

Also included in this repository is a collection of previous entries to the GPPC - notably some A* implementations. These were used for emperical testing and also for data to be compiled into the paper being written on this topic.

This code was submitted to the GPPC 2015 as open source.