Home
en fr

Fleet Optimization

Finding the sweet spot to run business smoothly

Context

Our client is a startup specialized in short-term car renting (from 1 to 30 days) and operates as a B2C company. To gain shares in a well established market, our client offers high end cars and above all a completely new digitized experience in order to avoid long waiting lines at a rental car desk. After an initial stage where they focus deeply on their growth, our client now wants to stabilize the current business and starts making benefits out of the oldest station (a dozen in France at the time), by allocating the right amount of cars for each station. Indeed, a high number of cars at a station helps to grasp the peak periods but has a strong impact on costs (parking costs and maintenance mainly). Conversely, a small amount of cars can create a regular shortage and therefore impact negatively the customer satisfaction on top of revenue leakage. The total number of cars in the park must be decided once a year as per contract with the car manufacturer. This huge decision lets few space for error and too little agility to absorb activity peaks for certain stations subject to higher volatility between high and low seasons. MFG Labs has been consulted to run analysis and model the business for each station in order to find the optimal number of cars that would maximize the profitability of the given station

Fleet Optimization

Solution

As a first step, we had to get a clear view of the business and how we could help. We led a serie of interviews with the top management to understand the main components of the business (operations, pricing, sourcing, product), the key decisions that were taken at each step and the main factor that could influence them. It helped us narrowing down what would be the main objective of the mission : find the right amount of vehicles for each station while our client first contacted us to help them with their pricing. We could then orient the problem solving towards a problem of optimization under constraints.

We decided to model our client’s business with an objective function for each station. The entry parameters were the number of cars to allocate over the year, and how the cars get allocated to the station during the year (a car must be hold by our client during a period of 6 to 18 months, so our client can hold a car from March to October for instance, not only from January till December). The function would output a benefit.

We applied an iterative mathematical approach, mixing simulation and optimisation techniques, to determine for each station an ideal number of cars and a distribution of them over the year that would maximize the benefit for each station. distribution Exemple de simulation / optimisation

Results

Our modeling presented certain limitation as we used some heuristics and used a known demand. However, it helped the top management to gain huge insights on their business and operations.

Beyond optimization, whose object is to provide an ideal distribution of the vehicles to hold under certain conditions, our approach based on simulations helped our clients to understand the non-linear characteristics of their business, the dynamic of vehicle holding, mostly for the station with high variability in demand for which the prediction are the hardest. It also helped them understand the impact on the P&L of over-stocking cars in stations where the demand is constant through the year.

The results, sometimes counter-intuitive, help them highlight some under estimated part of their business in order to better prepare their future.