Generating Initial Units in Games Using Evolutionary Algorithms and Dominance Relationship Matrix

The design of initial units in Real-Time Strategy games plays a crucial role in the gaming experience, providing players with a framework for early development and influencing strategies and decisions throughout the game. As RTS games evolve, the design of initial units continually innovates to accommodate the needs of players and the complexity of the game. Traditional resource allocation methods often rely on experience and intuition, lacking systematic and scientific approaches. However, designing these initial configurations is a time-consuming and labor-intensive process, requiring a significant amount of time and resources from game designers. Therefore, Procedural Content Generation (PCG) technology presents an attractive solution to this challenge. This paper proposes a novel PCG technique to expedite this process. Our method utilizes a comprehensive unit attribute system to construct a dominance relationship matrix, followed by the application of a genetic algorithm to optimize the initial units. We evaluate our technique in the scenario of generating initial units for real-time strategy games, resulting in balanced and diverse initial units.