BEST Center

MRL Building
Hastings Road
University Park, PA 16802
(814) 865-6237


BEST - Research: Tao Yao

Tao Yao

My research interests include optimization, stochastic models, game theory and their applications in transportation, energy, real options, outsourcing and supply chain management, revenue management, and financial engineering. In particular, I'm interested in development of cutting-edge optimization techniques such as robust optimization and stochastic optimization, and modeling and algorithm design for game-theoretic mathematical programs. These mathematical tools have been applied to the following projects:

Eco-routing for electric vehicles: “Eco-routing”, which aims at finding the most ecological transportation route, is attracting more and more attention from academia and industry due to the rising price of fuel and the mission of achieving environmental sustainability (see e.g., UTC Grant Solicitation, 2011) in nowadays. Recent developments in alternative and clean energy for vehicles (e.g. electric vehicles (EVs)) have brought new challenging and opportunity to this topic. In particular, due to the unique characteristics of EVs (e.g. limited battery capacity, ability of regenerating energy from wind and solar power, and design of regenerative braking), eco-routing for EVs is different and more important than other regular routing schemes (e.g. shortest path, fastest path, etc.). We aim to study eco-routing for EVs in order to make more efficient use of the new sustainable technology.

Traffic flow modeling, estimation, and control: We investigate fundamental questions regarding the modeling, estimation, and network design under uncertainty for complex transportation service systems. The rapid development of mobile internet (such as GPS-equipped vehicle and smartphone) has significantly improved the capabilities of the engineering community to monitor and control traffic. Importantly, mobile sensing has brought many opportunities and challenges for large-scale infrastructure systems. We propose tractable and scalable approaches (front tracking and robust optimization) to address traffic dynamics and uncertainties and to directly handle mobile sensing data.

Evacuation transportation planning: Large-scale evacuation in events is of critical importance since it bears significant infeasibility cost resulting from the potential loss of life and property. It is challenging to model the problem mathematically due to the inherent complexity and uncertainty. We used cell transmission model (CTM) based dynamic traffic assignment (DTA) to model the evacuation transportation problem and applied robust optimization and chance-constrained optimization to model demand uncertainty. Tractable formulations were developed and numerical experiments were conducted. Result showed that our approach leads to less total social cost than the deterministic approach. We plan to extend the study by considering more sources of uncertainty (e.g. road capacity uncertainty, evacuees' behavior uncertainty).

Congestion pricing and congestion derivatives: Congestion caused extra travel of 4.8 billion hours in US in 2009. This accounts an extra 3.9 billion gallons of fuel and a $115 billion congestion cost according to "2010 Urban Mobility report". There have been various attempts to mitigate congestion phenomena by increasing physical capacity, managing traffic demand and inducing alternative routing. Among them, Congestion pricing is widely accepted as an effective way and has already been applied not only to highways but also urban roads. We developed robust dynamic game-theoretic congestion pricing model which addressed several challenges such as uncertainty, competition between users and dynamic price, and designed advanced heuristic algorithm to solve it. In addition to the congestion pricing, we also proposed congestion derivatives for reducing congestion and the related total social cost. Theoretical analysis and numerical tests were conducted to prove the value of congestion derivatives. We plan to develop less conservative model and consider more general uncertainty sets for congestion pricing, and investigate more general network models as well as various types of securities for the study of congestion derivatives.

Urban freight transportation:Growing demand for urban freight transportation services and the recognition of the value of sustainable transport make efficiency and ecology the two main issues for the urban freight transportation system of the 21st century. Considering that the private vehicles transporting people are also a main component of the urban transportation system and their interactions with vehicles transporting freight were often ignored in the study, we investigated their interactions by formulating the urban freight transportation model as a mathematical program with equilibrium constraints (MPEC) and develop accurate measures of the interactions. We solved the problem numerically and discussed some managerial insights for truck companies and central planner. The study will be extended by modeling freight transportation as vehicle routing problems and considering more realistic issues like policies, restrictions and uncertainties.

Complex transportation network design: Recently, we proposed to develop a new complex network design theory which is dynamic and considers the influence of transportation, social, and data networks on one another. In particular, we will consider multiple time scales and human behaviors when modeling traffic flows and demand formation. Network science and evolutionary game theory will be used to model demand dynamics. Dynamic arc and path flows will be modeled based on hydrodynamic traffic flow theory.