The project FITS aims at proposing novel mathematical models and solution approaches in order to develop intelligent tools to help optimizing the delivery of services (such as care or maintenance deliveries) to customers located potentially at home. Service providers have human resources (possibly outsourced) with different characteristics. They can have different skills (specialized to versatile), different availabilities and different locations. Customer demands can be erratic, dynamic, stochastic and heterogeneous.
The goal of a service provider is to assign customers’ demands to human resources in order to ensure a high quality of service while maintaining good working conditions. For example, in order to satisfy human resources, their activity should be diversified and the driving distance minimized; and in order to satisfy customers, service delivery should be regular and consistent.
In order to ensure such an assignment, there exist reservation platforms that aim at putting in touch human resources and customers (servilink, allovoisin, domicalis). That business met recent success stories: Airbnb, Blablacar and Uber. This success shows that an increasing number of customers are web-connected and they require more than responsive services as they expect that some future needs have to be anticipated. The efficiency of these systems strongly depends on the quality of data, and supply and demand matching. Capturing and storing data becomes easier and easier, but extracting knowledge from these data and using these data efficiently remains a difficult task.
When both demand and offer reach high volumes, some basic rules may satisfy users. But several cases can make these systems less efficient. Some of these cases are: (1) Low demand density, (2) Low supply offer, (3) Strong constraints, and (4) Strong demand variability.
Thus, if smart systems are not available to deal with these problems, some low density population areas and some delicate or strictly constrained activities cannot take advantage of service platforms. So several regions and activities may wonder whether they can or cannot apply these services. In order to ensure a good quality of service while maintaining good working conditions and a profitable system, such platforms need to be optimized.
The project FITS is designed to tackle the scientific challenges raised by the optimization of these service provider platforms that need intelligent tools for responsive large scale transportation services. FITS is designed to address those issues by taking profits from available data (real-time information and statistics on large data). FITS has to perform a smart assignment between users in order to find the best balance between satisfying customers and covering constraining or less profitable requests with fair dispatching rules among workers in terms of difficulty and profit.
FITS is conducted by a consortium of three complementary teams (CMP, CIS and UCA) in computer sciences, operations research, management science and healthcare engineering. Developed algorithms will be tested on real data extracted from the open living lab #futuremedicine MedTechDesign. All developed algorithms will be prototyped in this living lab.
Monsieur Nabil Absi (Mines Saint-Etienne - CMP)
The author of this summary is the project coordinator, who is responsible for the content of this summary. The ANR declines any responsibility as for its contents.
CMP/EMSE Mines Saint-Etienne - CMP
CIS/EMSE Mines Saint-Etienne - CIS
UCA/LIMOS Université Clermont Auvergne/LIMOS
Help of the ANR 465,480 euros
Beginning and duration of the scientific project: January 2019 - 48 Months