Ride-pooling optimization typically concerns with matching, repositioning, routing (see e.g., Zheng et al. (2018); Alonso-Mora et al. (2017a, 2017b); Tong et al. (2018)). The RL literature has primarily focused on the first two problems. The ride-pooling matching problem differs from that in Chap. 5 in that a combination of multiple passengers, and hence their combined trip, can be matched to a vehicle that may or may not be empty. See stages B and C in Fig. 8.1 from Alonso-Mora et al. (2017a) for an illustration. The repositioning problem is similar to the ride-hailing case, except that the objective is to optimize some pooling-specific metrics that we define next. The routing problem solves for the sequence of pick-ups and drop-offs given the assigned passengers for a vehicle. The routing problem could also concern with route guidance on the road network. See stage D in Fig. 8.1.