Topology-based generation of sport training sessions

Recently, sports training sessions have been generated automatically according to the TRIMP load quantifier that can be calculated easily using data obtained from mobile devices worn by an athlete during the session. This paper focuses on generating a sport training session in cycling, and bases on data obtained from power-meters that, nowadays, present unavoidable tools for cyclists. In line with this, the TSS load quantifier, based on power-meter data, was applied, while the training plan was constructed from a topology of already realized training sessions represented as a topological graph, where the edges in the graph are equipped with the real length, absolute ascent and average power needed for overcoming the path between incident nodes. The problem is defined as an optimization, where the optimal path between two user selected nodes is searched for, and solved with an Evolutionary Algorithm using variable length representation of individuals, an evaluation function inspired by the TSS quantifier, while the variation operators must be adjusted to work with the representation. The results, performed on an archive of sports training sessions by an amateur cyclist showed the suitability of the method also in practice.