Individualized optimal strategy in racing sports: the case study of the team pursuit in track cycling.

Abstract

In track cycling, performance in the team pursuit depends on the mechanical and physiological abilities of each member of the team, but also on the choice of racing strategy. Athletes must cover the 4000 m of the race, sharing the effort between them in successive relays. This raises the question of the optimum strategy. We propose a method for resolving this question by coupling a mechanical model of the race to physiological models (digital twins) of the athletes. The mechanical model enables one to predict a theoretical finishing time for a given strategy, while the physiological model enables one to determine whether or not a given strategy is feasible. By coupling the two models and using numerical optimization, an optimal strategy for a given team can then be predicted. Simplified team composition case studies are explored. For each case studied, an optimal strategy to maximize performance is obtained and composed of a set of three variables: relay lengths, power values for each relay, and the starting order of cyclists. The proposed method can be used for real athletes and extended to other disciplines.