ContMulti-objective Optimization Model for Momentum Change Based on Genetic Algorithm

This paper uses key events in the dataset as indicators of player momentum changes to explore the dynamic changes in momentum during matches. Using genetic algorithm to optimize the momentum change amplitude of different players in key events, forming a multi-objective optimization model. Firstly, process the data and fill in missing values, and test the correlation between the data through correlation analysis. Select key events that affect player status as momentum change events and use them as optimization objectives for genetic algorithms. The accuracy of momentum prediction for the next serve winning rate is about 60%. The cointegration test and Granger causality analysis determined the impact of momentum on game scores. Using the established optimization model, an analysis of a game was conducted, and the results showed that the score and penalty were the most critical factors. A universality test was conducted on a table tennis match, with an accuracy of approximately 65%, but it was affected by a lack of dataset and information. There is a covariance between the score difference and the player's momentum change in the game, but there is no clear causal relationship.