This paper explores optimising team line-ups in wheelchair basketball, a sport governed by the International Wheelchair Basketball Federation (IWBF) and its Player Classification Points System (PCPS). It focuses on evaluating players' performances using metrics like the efficiency rating (EFF), performance index rating (PIR), and Win score. The study employs Bayesian longitudinal modelling alongside an optimisation programming respecting game constraints, including players' gender, classification, and match location, to predict the best team compositions. This methodology is applied to the Doneck Dolphins Trier team from the German Rollstuhlbasketball-Bundesliga (RBB), highlighting the effectiveness of these models in preparing for the 2022-2023 season matches. The study demonstrates a novel approach to team optimisation in wheelchair basketball, considering both player performance and game regulations.