NSGA-II for city building placement optimization in the turn-based game civilization VI

Building a good city in the video game Civilization VI is important for the players. It consists of a series of choosing what to build and where to place buildings in the city to generate high resources for faster pace in the race of winning the game. In this paper, we try to create an agent that able to plan a city building by finding optimized combinations of building placements to get high and balanced yields for the city. There are six yields that can be generated by a city and each yield has importance on its own. We propose solving this multiobjective optimization problem using Multiobjective Evolutionary Optimization (MOEA) Non-dominated Sorting Genetic Algorithm II (NSGA II). NSGA II provides pareto fronts values. We compare a city built by our simulation with a city built by native in-game AI. Unlike the in-game AI which is always biased to build farm improvements resulting yields that are dominated by food yield, our simulation are able to generate a selection of building placements solution with both high yield and more yield diversity.