In the following we present the PortLib instances for the Port Scheduling Problem (PSP), which have been presented in the paper An Adaptive Large Neighbourhood Search Heuristic for Routing and Scheduling Feeder Vessels in Multi-terminal Ports, written by Erik Hellsten, David Sacramento and David Pisinger, and published in European Journal of Operational Research. The repository includes the results for PortLib instances for the Adaptive Large Neighbourhood Search (ALNS) heuristic and the commercial solver CPLEX. Additionally, we further include the results for the Constraint Programming and the ALNS Math-heuristic approaches from the paper Constraint Programming and Local Search Heuristic: A Math-heuristic Approach for Routing and Scheduling Feeder Vessels in Multi-Terminal Ports, written by David Sacramento, Christine Solnon and David Pisinger, and pending for publication in SN Operations Research Forum. Furthermore, this version includes the results for the Constraint Programming models from the paper Integrated Planning of Feeder Vessels at Multi-Terminal Ports, written by David Sacramento and David Pisinger, and pending for publication in 4OR - A Quarterly Journal of Operations Research. The PSP represents a new scheduling problem for feeder vessels in multi-terminal ports, which has been defined in close collaboration with the industry. The proposed problem is a General Shop-like problem, and it accounts for most of the practical restrictions faced by the carriers in scheduling the operations. Given a fleet of feeder vessels, which each of them has a number of operations to perform at different terminals, and each terminal can only serve one vessels at a time, the task is to define an operational schedule, i.e. a starting time for each operation, which satisfies the time window and precedence constraints as well as minimises the departure times of the vessels and packs the schedule as tight as possible. The instances are named PSP.n.m.r, where n is the number of container-terminals, m is the number of vessels, and r is the generic name of the scenario. The instances are randomly generated to be realistic, but in addition we ensured that each instance has a feasible solution as well as strove towards that each constraint should have a significant impact. In general, the instances are made to be slightly harder to solve than the problems faced by industry, in order to properly challenge the developed methods, as well as spurring further development.