Abstract

As of 2018, Montreal’s bicycle-sharing system (BSS), BIXI Montreal, is managing about 6,200 bicycles and 540 stations from mid-April till mid-November. A major problem that arises in BSS’s is bicycle and/or dock availability. A more predictable system will help optimize bicycle-rebalancing and raise the customers’ satisfaction of the service which, eventually, will generate incentives for mode-shifting and reducing traffic congestion. In this thesis, we used the 2017 open-source data along with basic demographic information, provided by BIXI, to create a methodology that will improve rebalancing procedures and add adjustability to a BSS. Preliminary investigation of the data shows that bicycle/dock availabilities are critical during the AM peak hours - when most users utilize the system for commute. Arriving on time in the morning is a priority for most users and providing a convenient infrastructure service for this is believed to be essential for any growing city. At the station level, we examine arrivals and departures as well as capacity and inter-station distances. At the user level, we examine gender, age, language and trip duration. The developed methods will allow bicycle sharing systems to study the variation of the Time of Blockage (ToB) as the user profile changes. This is very useful when there is interest in geographical expansion of the system. The methods will also allow for real-time monitoring of the ToB instead of using pre-set dispatch times. This will definitely optimize bicycle rebalancing operations - maximizing convenience, availability and users’ satisfaction.