A Mixed-Integer Linear Program (MILP) is necessary in routing and scheduling scenarios where decisions involve both discrete choices (like selecting which vehicle, route, or time slot to use) and continuous measurements (like travel time, arrival timestamps, fuel consumption, or continuous load weights).

You need MILP whenever you have both yes/no decisions and values that can naturally be fractions, such as in scheduling where binary variables assign jobs to machines or select routes while continuous variables represent start times, production rates, or load sizes that do not need to be whole numbers; the integer variables set the overall structure—what is active or chosen—while the continuous variables fill in the details, and they are linked by rules like “if the binary choice is 0, the continuous value drops to 0,” so pure integer programming would unnecessarily restrict realistic quantities to integers and produce less accurate or useful results.

An example is the Capacitated Vehicle Routing Problem with Time Windows, where discrete decisions like assigning vehicles to customer routes require binary integer variables while arrival times and payload quantities require continuous variables. Continuous variables interact with integer variables through linking constraints that enforce operational limits, such as forcing a delivery window or truck load variable to zero whenever a route binary variable is inactive. This hybrid interaction allows the model to enforce logical routing paths while simultaneously optimizing continuous physical quantities like fuel, time, and inventory.

A Mixed-Integer Linear Program (MILP) is necessary in scheduling or routing scenarios that require balancing discrete operational choices with continuous physical constraints—such as the Vehicle Routing Problem with Time Windows and Fuel/Payload Consumption, or power plant generation scheduling. In these applications, integer (or binary) variables handle discrete logic, such as deciding whether a truck serves a specific customer (1 or 0) or which sequence of jobs a machine processes, while continuous variables track ongoing, non-discrete quantities like exact arrival times, fuel levels, remaining driver shift hours, or continuous power outputs. The continuous and integer variables interact through linking constraints (often using "big-M" formulations), where a binary variable acts as an on/off switch that conditionally activates, enforces, or limits continuous bounds—for instance, forcing a vehicle's departure time variable to stay at zero unless the binary variable for that route is activated, or capping a plant's continuous output level strictly within its operational limits only when the plant is switched on.