Using Graph Theory to Investigate the Role of Expertise on Infrastructure Evolution: A Case Study Examining the Game Factorio

Research into critical infrastructure network architecture design faces two significant challenges. First, real-world network performance data is often not available due to being proprietary. Second, many efforts focus on analyzing the structure of an infrastructure network at a single point in time, while real-world networks are constantly evolving. In this chapter, these two gaps (need for more data and for time-series data) are examined by utilizing a new data source: the video game Factorio. Factorio is a manufacturing simulator. Utilizing publicly available recordings of players’ networks in game, a shared end point, and completion time stamps allows the examination of different network strategies. The key research question examined in this work is how does network evolution change when comparing ten expert and ten novice designers? This chapter provides two key contributions. First, a qualitative and quantitative analysis of how ten different structural graph theory metrics evolve when comparing expert and novice designers is provided. The expert dataset has a narrower distribution, indicating common strategies, and focuses on critical path manufacturing early in the network’s evolution. The second contribution is a set of time-series network data that can be used for additional studies. By examining the differences in network evolution between experts and novices, this article performs a critical first step toward using in situ graph theory metrics as a decision aid for designers during infrastructure evolution.