Build Strategic Alliance Model for Playing Empire Kingdom Video Game

摘要

The objective of this paper is to introduce a novel methodology of playing a strategic video game by applying several statistical modeling techniques to meet the required troop quality metric. The focus is on how to build a powerful troop which can both attack the remote, stronger castles and could also defend the main castles when attacking troops were far away. K-Means Clustering computing was utilized to cluster the nearby castles to form a joint alliance which was able to communicate and support each other against any mighty attacking from nearby Kingdoms. Principal Component Analysis (PCA) was utilized to study 7 troop characteristics among 40+ troop types available. The PCA Eigenvalue and Eigenvector analysis could effectively provide the insights of troop improvement strategy based on the PCA patterns and each principal component. Established the Transfer Function Sensitivity Analysis through Custom Profile to conduct the troop design optimization. Based on provided Kingdom's expansion strategy, the optimal troop design was identified. The design optimization algorithm and the optimal design were fully aligned or predicted by PCA eigen analysis. This PCA-based algorithm could maximize the Return of Investment (RDI) on playing strategic video game. The same PCA modeling technique could be extended to most modern Business Management World.