Real-world applications often require the resource-specific configuration of process instances based on system states. In clinical processes, for example, the number and order of documentation tasks depend on whether they are conducted by an intern or the head physician. One approach supporting resource-specific instance configuration is the Resource-Augmented Process Structure Tree (RA-PST). It also offers the flexibility to drive the configuration through optimized resource allocation. In particular, combinatorial optimization approaches can be explored to plan optimal combinations of multiple configurable process instances into a system. We use Mixed Integer Programming (MIP) to find optimal process configurations and Constraint Programming (CP) to schedule the process instances. To combine configuration and scheduling of process instances, we propose an integrated CP formulation as well as a Logic-Based Benders Decomposition (LBBD) inspired by the integrated process planning and scheduling problem (IPPS). The approaches are applied and evaluated in an offline and an online scheduling setting. We show the suitability of the CP and LBBD formulation for the configuration and scheduling problem. The integrated CP can find optimal solutions for small problems but struggles to prove optimality in a timely manner when the problem size grows. Here, the LBBD can identify better lower bounds. For the online setting, the integrated CP finds optimal solutions for most process instances while considering the planned resource availabilities in the system.