
Regarding a linear programming method for approaching this problem, the critical path method was the original idea. The goal of this project was to apply linear programming to their notion of optimizing build orders and assess its efficacy. Methods included a depth-first branch and bound algorithm that takes a game state and performs a recursive search on possible build orders that satisfy a particular goal. The authors successfully created a “planner” that when incorporated with a playing agent performed comparable to professionals. Inspiration comes from Churchill and Buro’s Build Order Optimization in StarCraft, an article focused on artificial intelligence advancement with respect to StarCraft.


A player begins with nothing but a base and six worker units meant for collecting resources (minerals and gas) that are used to produce structures and more advanced combat and support units.

A player has a choice of playing as three different races – Terran, Protoss, or Zerg – all with unique units, buildings, abilities, and mechanics. Essentially a war simulation, the goal of the game is to eliminate your opponent by amassing an army and making your opponent forfeit, or destroying all of your opponent’s structures. StarCraft launched in 1998 and gave rise to celebrity-status professionals and prestigious tournaments, and is still updated with new content and expansions. The StarCraft franchise developed by Blizzard Entertainment is one of the most popular real-time strategy video games. A seventh is being created in the Command Center. A Terran base with the starting six workers.
