Interactive Operations Research

Parking Facility Design for Autonomous Vehicles 

Autonomous vehicles will have a major impact on parking facility designs in the future. Compared to regular car-parks that have only two rows of vehicles in each island, future car-parks (for autonomous vehicles) can have multiple rows of vehicles stacked behind each other. Although this multi-row layout reduces parking space, it can cause blockage if a certain vehicle is barricaded by other vehicles and cannot leave the facility. To release barricaded vehicles, the car-park operator has to relocate some of the vehicles to create a clear pathway for the blocked vehicle to exit. The extent of vehicle relocation depends on the layout design of the car-park. Our research shows that autonomous vehicle car-parks can decrease the need for parking space by an average of 62% and a maximum of 87%. This revitalization of space that was previously used for parking can be socially beneficial if car-parks are converted into commercial and residential land-uses.

Loyalty Program in Public and Private Transportation

The proliferation of smart cards in public transportation has paved the way for successful implementation of two prominent discount policies: pass programs and loyalty programs. While pass programs have been around as early as the 1970s, loyalty programs are only now gaining unprecedented popularity in public transportation. In a loyalty program, riders get a discount on their fare if they complete a given number of trips within a time period (e.g. a month). For example, Presto offers a 11% discount to riders that make more than 35 trips in one month. 

Our recent research shows that public transit agencies are better off with a pass program to maximize social welfare whereas private agencies like Uber and Lyft should offer the loyalty program to maximize profit. In addition to these, we also found the optimal design of the loyalty program, i.e., how much discount should be provided after how many trips. 

Dynamic Taxi Dispatching with Congestion Effects 

Taxis are increasingly becoming a prominent mobility mode in many major cities due to their accessibility and convenience. The growing number of taxi trips and the increasing contribution of taxis to traffic congestion are cause for concern when vacant taxis are not distributed optimally within the city and are unable to find unserved passengers
effectively. A way of improving taxi operations is to deploy a taxi dispatch system that matches the vacant taxis and waiting passengers while considering the search friction dynamics.

We developed an optimal taxi dispatch model and tested it on a two-region city (see video).  The results show that lack of taxi dispatching leads to severe accumulation of unserved passengers and vacant taxis in different regions whereas the dispatch system improves the taxi service performance and reduces traffic congestion. The proposed framework demonstrates sound potential management schemes for emerging mobility solutions such as fleet of automated vehicles and demand-responsive transit services.

Simulating Terminal 3 Curbside of Pearson International Airport in Toronto

Airport curbside congestion is a growing problem as airport passenger traffic continues to increase. Many airports accommodate the increase in passenger traffic by relying on policy and design measures to alleviate congestion and optimize operations. We simulated scenarios such as double parking, alternative parking space allocation, increased passenger demand, and enforced dwell times at Pearson International Airport in Toronto, Canada. The results show that double parking reduces the utilization ratio and the level-of-service of the outer curbside but cuts down the passenger and vehicle waiting time. Inclement weather conditions reduce the utilization ratio of the inner curbside and the supply of commercial vehicles since it takes them longer to return to the airport. Finally, reducing the allowable parking time at the curbside decreases the average dwell time of private vehicles from 89 seconds to 75 seconds but increases the number of circulating vehicles by 30%.

Impact of Hourly Parking Pricing 

Hourly parking pricing is a common parking management strategy where vehicles pay based on their parking duration (dwell time). Although pricing is inherently assumed to reduce demand (for parking), this is not the case when we have hourly parking pricing, i.e., the demand may increase or decrease with the price. Hence, hourly pricing can actually cause higher congestion and decay social welfare if imposed imprudently.

Continuous Approximation of the Fleet Composition Problem

Distribution companies in the freight industry own a fleet of vehicles (trucks) to serve a given demand. Finding the optimal fleet composition (i.e. how many and what type of vehicles to purchase) is a complex problem to solve to optimality for large instances. Here, we present a methodology that divides the service region into sub-regions, each assigned to one vehicle. This methodology provides insights that are otherwise harder to gain from conventional methods. 

Network learning via multi-agent inverse transportation problems

Existing methods of (behavior-based) travel time estimation for a network are often incompatible with ubiquitous data that arrive continually. In other words, it is cumbersome to continuously feed (estimate and calibrate) existing models with real-time data. To address this issue, we developed a model that estimates network travel times by learning the route choice behavior of travellers. Here's how the model works. We assume travellers choose the shortest (optimal) route when going from point A to point B according to the network travel times. Obviously, with one observation, we cannot say much about the travel times. However, as the number of travellers increases, we can more accurately infer network travel times. The methodology is based on Inverse Optimization. The model was successfully tested on a number of networks including a freeway network in Queens, New York City (See figure). Click here to see more. 

A dynamic carsharing decision support system

Carhsaring services have become popular in the last decade due growing interests in the shared economy. One challenge in managing a carsharing service is the imbalance of vehicles in different stations which arises from the demand pattern of the users; we could have too many cars in one station and no cars in another. The imbalance issue can be resolved with the arrival of autonomous vehicles as these vehicles can be self-relocated between stations to bring the system back to balance. We studied the optimal vehicle relocation in the downtown area of Toronto. We found an inverse relationship between the fleet size and the total relocation time. When the fleet size is large, there is little need for relocations. Smaller fleets, on the other hand, require more relocation (See figure). 

Vehicle-to-grid Technology

Vehicle-to-grid (V2G) is a technology which can reduce the cost for power distribution network operators by storing electricity in the batteries of electric-drive vehicles and retrieving it when energy demands increase during the course of a day. Participants of V2G are reimbursed for offering their vehicles which can lead to changes in trip schedules when V2G payments are high and travelers are sensitive to the payments. However, prior studies have ignored the effects of V2G on travelers’ schedules. 

Impacts of Autonomous Vehicles on Cities 

Autonomous vehicles are predicted to enter the consumer market in less than a decade. However, there is no consensus on whether their presence will have a positive impact on users and society. The skeptics of automation foresee increased congestion, whereas the advocates envision smoother traffic with shorter travel times. In this paper, we study the automation controversy using supply-demand analysis. We derive several managerial insights. First, we show that a sound judgment of automation relies on the occurrence of three possible cases for which full, null, or partial automation is recommended. Second, although traffic increases with automation, the travel times may decrease. Third, we show that autonomous vehicle owners travel more than regular vehicle owners because they can engage in alternative activities (e.g., reading) while in the vehicle. We also propose three policies to promote automation when it is deemed beneficial. The three policies are based on subsidization, taxation-and-subsidization (under revenue neutrality), and vehicle-sharing. We show that subsidization leads to higher social welfare than tax-and-subsidize, but a higher level of automation is achieved by the latter. We also show that the optimal policy depends the ability of the infrastructure to serve autonomous vehicles, hence, the optimal policy may change with time as the infrastructure is improved.