Technology, Integration and Scalability
Optimal Outputs uses advanced predictive analytics, simulation and machine learning techniques within our highly scalable cloud based platform to analyse, predict and recommend the best fleet for your business.
OptimalCar is developed in Java by a highly experienced development team comprising Software Architects and Engineers, Operations Research specialists and experienced Car Rental Fleet Planning and Revenue Management Professionals.
Technology
OptimalCar has been developed to handle the specific complexities of the car rental industry. Car Rental is complex and producing accurate projections and recommendations requires advanced technologies and techniques:
OptimalCar uses Monte Carlo Simulation techniques to increase the accuracy of our projections. Each Projection involves running up to 1000 simulations for each station for the next 36 months to give long term visibility on fleet requirements and profitability. Machine Learning processes are used to continually improve the accuracy of certain simulation inputs.
Integration
OptimalCar integrates easily with existing Car Rental Systems. In addition to standard data such as Car Classes and Categories, Rate Codes and Station Information, Optimal Car uses historical and current fleet, rental agreement and future booking data.
Data updates containing the day's update are transferred to OptimalCar using File Transfer or Email attachments. This data is automatically imported into the system and a new simulation process is started to update the forecasts. Where the car rental company prefers to use another forecasting system, forecast data can be imported and OptimalCar will run simulations and recommend fleet changes based on that forecast data.
Scalability
OptimalCar performs 'bottom up' simulation of every rental location in your network. Projections for all locations are then aggregated to give overall visibility across the network.
OptimalCar runs on Amazon Web Services Cloud Compute and Database services using 'on demand' compute power with distributed processing to run projections as needed. The system can automatically scale to support the largest numbers of vehicles and stations by expanding the number of our projection servers working in parallel.