A Matching Framework for Employees to Support Carpooling in the Context of Large Companies

Abstract

Matching potential carpool partners in large companies is one of the critical tools to establish carpooling. The main practical problem of every matching framework is that it only starts working well when a sufficiently large set of candidates is available. For this reason, public carpool matchers have fairly low success. This paper presents the components and development decisions for a closed group framework for matching employees who are candidates for carpooling. It is designed to be operated by employers in order to find optimal carpool matching solutions which are to be proposed to candidate carpoolers. It has the capability to account for dynamic evolution of the extracted personnel database in order to minimize burden on the users. This framework is capable to match candidates based on their home and target locations, time windows, allowed detour durations and several attributes describing personal behavioral properties specified by the interested candidates. People fix their choice after negotiation within small groups and feedback their decision to the system that maintains a carpool calendar and personal preferences for every participant. The result is a dynamic system that evolves to a user optimum (as opposed to system optimum) and therefore can be considered as stable. As a proof-of-concept, experiments were conducted at the scale of the Doppahuis database. The matching framework is sufficiently efficient to recompute the advice to the customers after changes in the carpooling candidates' database and due to unexpected changes in daily travel requirements. Results show that the computation time of the framework grows in a polynomial way with the scale of the potential carpoolers in a carpool.