New fleet management solution

Kurzfassung

An SME located in Southern Italy (Calabria Region) operating in ICT sector, has developed an innovative system (hardware and software) for real time fleet management. The solution is based on the integration of innovative forecasting/optimization methodologies and technologies such as: GPRS, GPS and GIS. It is mainly addressed to companies operating in urban freight delivery sector (same-day delivery). The company is looking for technical and/ or commercial collaboration.

Details

An SME located in Southern Italy (Calabria Region) operating in ICT sector, has developed an innovative system (hardware and software) for real time fleet management. It is mainly addressed to companies operating in urban freight delivery sector (same day delivery), dial-a-ride transportation systems, rescue and repair services, taxi cab services as well as emergency services (police, fire fighting and ambulance services).

The system operates by a software tool, to be tactically used to determine the optimal fleet composition (i.e., the number and type of vehicles and crews to be hired at any moment) as a function of the period of the year, the day of the week, the period of the day, etc.

The system is based on the integration of different technologies: GPS/GPRS (each courier has got a GPRS Personal Digital Assistant with GPS receiver), GIS (for internet connection on Personal Digital Assistant and to trace vehicles from the headquarter), optimization techniques to route the vehicles, forecasting techniques based on neutral network, parallel computing. Each courier is connected to the headquarter by a specific software installed on the Personal Digitial Assistant communicating via GPRS: its position is traced using the data coming from GPS receivers and GIS data. The duration of any single trip is estimated by a neutral network taking into account the type of vehicle, the meteorological and traffic data. The allocation system automatically chooses the most convenient courier to serve a particular request on the basis of the following elements: type of vehicle requested; level of service foreseen for the client (SLA “service level agreement”), localisation of the courier, the requests currently delivered by each courier, the weather conditions and traffic situation. For each courier the system defines the optimal route to deliver, minimising the length of the route.

Furthermore, the system guarantees the automatic reposition of the “unloaded” vehicles taking into account the urban areas with high probability to be involved by further requests.

Innovative Aspects:
Innovation
The traditional model of service utilises human controllers who ask vehicles to relay location information and then assign jobs on the basis of distance. This model is not only inefficient, but also suffers from errors inherent with the involvement of human factors. As the number of vehicles increases to several hundreds or thousands, complexity grows to levels which push the limits of human analysis.

There are also several software packages which allow to route and schedule, but they assume demand is known in advance (e.g., route plans are generated daily on the basis of the customer requests received the day before) or moderately dynamic. This case includes newspaper delivery, refuelling gas stations, postal delivery and garbage collection, to name a few.

The solution goes over these weaknesses using innovative on-line forecasting/optimization algorithms operating at two different levels:
• OPERATIONAL LEVEL

- Real-time forecasting algorithms:
An innovative methodologies was developed in which historical data forecasts is wisely combined with real-time info provided by the whole fleet. The basic idea is that the travel time of a segment of route to be traversed in the near future is positively correlated, from a statistical point of view, to the travel times experienced by other vehicles which have traversed the same “area” during the “near past”.

- Anticipatory dispatching algorithms:
The current dispatching algorithms react to new requests only once they have occurred, while neglecting available stochastic information. To overtake such a situation the company exploited two different strategies:
a) first, it developed an innovative anticipatory algorithms that evaluate alternative solutions through a short-term demand sampling followed by a fully sequential procedure for indifference zone selection.;
b) second, the most relevant demand patterns was extracted through data mining techniques and devised fast anticipatory rules addressing those patterns.

• TACTICAL LEVEL

- Shift scheduling algorithms:
Real-time fleet management poses the problem of minimizing the overall staffing cost subject to service-level requirements. The problem was modelled as an integer program with non linear probabilistic constraints and solved through simulation-optimization procedures. A neighbourhood evaluation (ANE) mechanisms was exploited based on the solution of deterministic optimization problems with a reduced number of parameters estimated via simulation. These shift scheduling algorithms provide significant cost reductions in a typical same-day courier setting.

Advantages
- reduction of response time in emergency systems (police, fire fighting, ambulance,etc which could help saving lives and increasing homeland security;
- reduction in traffic congestion, air pollution, energy consumption and safety.