Hi Peter,
With all orders being assigned to a specific driver that means there is no actual overlap between the drivers on the VRP and so you can actually make your problem simpler and be able to plan for the whole month. The first step will be to separate the problem into a different VRP model per driver so only the orders that are assigned for that driver are in the specific model and only that driver's routes are in the model. You can use the VRP layer's name to signify what driver that model is for and then set up 15 different layers for your 15 different drivers. This means you do not need to use specialties to signify the drivers since the matching has already been done.
The next step is to set up the routes so that there is one for each day in the month that you are wanting to model. It sounds like the work days for the drivers are the same settings so the routes can be copies of each other. Just change the naming so that it signifies the day that route will be taken such as Day1, Day2, Day3... You therefore do not need to use the breaks to signify the end of the day (you could use them for lunch breaks if needed) and each day is modeled as a different route.
Without any day of the week constraints on the orders, you are mostly just wanting the VRP solver to break the work up into different routes to use. I might suggest adding dynamic seed points to get some better clustering allowing the driver to stay in one area of the region on a specific day.
Let me know if you have more questions or need clarification on this.