Multiple Vehicles, Multiple Shifts
Subsequent examples focus on multiple vehicles servicing orders over a longer planning horizon consisting of multiple shifts (typically a shift is a day but it can be longer if desired). Each vehicle can have its own shifts so that vehicles can be operating during the same planning period or different periods. This can be useful when different vehicles are available on different days due to vehicle maintenance or employee availability.
When routing across multiple days in multiple vehicles, you may often have some orders that require periodic visits. Rather than specifying a separate order for each visit, the optimization engine allows you to specify a single order that has requirements for multiple visits. The
visit_range
constraint allows you to ensure the correct number of visits across the planning period. Additionally, since orders should generally be spaced out over the planning period, the
visit_gap
constraint provides control over how much time separates one visit from the next. These constraints operate at the order level regardless of which vehicle visits the order.
Back to our example of Jimmy and Sally’s expanding business in the DC Metro Area, based on their great success in matching the right worker for the right job at the embassies, they have now obtained contracts for cleaning the embassies along with the other work. The various embassies have different requirements for how often the service should occur. They will operate on a weekly cycle with the visit count requirements and visit spacing requirements now incorporated into the orders
ORDERS —
All of the orders now receive a value of min_visits – the minimum number of times we must visit them in the solution to the routing problem. This is enforced via the
visit_range
constraint. Since all these orders must now be visited multiple times, the desired spacing of visits is expressed with min_days and max_days – the minimum and the maximum number of days between service events. This spacing is enforced via the
visit_gap
constraint. Also, note that we add a few additional locations and orders to make sure the fleet has enough work to remain busy on most days.
CONSTRAINTS —
Since the
visit_range
is present in all examples to achieve the default min_visits of 1, the only addition is the
visit_gap
constraint. This constraint references the data present in each order object.
Expand to view request sample
Part of the solution for the multi-shift, multi-vehicle problem. We now have 15 total routes (3 vehicles, 5 days each), so it is difficult to make sense of all days and all vehicles at once. Each vehicle is active on all 5 days except for Vehicle 2 on Day 2 and Vehicle 3 on Day 4
Day 1 of the solution when all 3 vehicles are active.
A zoomed in view to show the travel inefficiencies that can result from enforcing the attribute matching. In this case, the green (Finnish speaking) vehicle services the green order #4 by going out of the way into the Embassy Row area. The orange and pink vehicles are nearby but do not match the attributes of the order at the Finnish Embassy, so they cannot service the order.
