Covering an array generation is one of the key issues in combinatorial testing, and algorithms are popular due to its ability to deliver smaller covering array in shorter time. People have proposed many greedy algorithms based on different strategies, and most of these can be integrated into a framework, which forms a configurable greedy algorithm. Many new algorithms can be developed within this framework, however, deploying and optimizing the framework affected by multiple factors to construct more efficient covering arrays is a new challenge. The paper designs three different experiments under the framework with six decisions, systematically explore the influence of each of the decisions and interactions among them, to find the best configuration for generating smaller covering array, and provide theoretical and practical guideline for the design and optimization of greedy algorithms.