[BUG] cuOpt VRP solver ignores min_vehicles constraint when fixed costs are included

When vehicle costs are included in the data VRP data model and min_vehicles constraint is set, cuOpt generates a solution with fewer than min_vehicles constraint value. Likely, this arises when min_vehicles value = total number of vehicles and fixed costs are included.

The expected behavior is for the solver to use no fewer than min_vehicles vehicles and pick the least costly vehicles as dictated by fixed costs.

The following example is where total num vehicles is 3, and they all have fixed costs + min_vehicles = 3, and the resulting solution uses only 1 vehicle.

{ "cost_matrix_data": { "data": { "0": [ [0, 17, 12, 11, 10, 18, 10], [16, 0, 15, 11, 19, 15, 16], [19, 19, 0, 11, 16, 11, 17], [17, 19, 17, 0, 11, 18, 19], [10, 19, 19, 19, 0, 17, 15], [12, 18, 15, 18, 18, 0, 14], [12, 12, 11, 19, 10, 17, 0] ] } }, "fleet_data": { "vehicle_locations": [ [0, 0], [0, 0], [0, 0] ], "vehicle_fixed_costs": [10, 20, 30], "min_vehicles": 3 }, "task_data": { "task_locations": [1, 2, 3, 4, 5, 6, 0, 0, 0, 0, 0, 0], "demand": [[2, 1, 4, 2, 1, 4, -2, -1, -4, -2, -1, -4]] }, "solver_config": { "objectives": { "cost": 1, "vehicle_fixed_cost": 1 } } }

When the vehicle_fixed_cost values are set to 0, the solution uses 3 vehicles, as required by min_vehicles setting.