New Approaches and Methodologies to Reduce Energy Consumption and Greenhouse Gas Emissions on Transportation Networks

The models and algorithms developed for transportation planning, delivery/pickup logistics, and vehicle routing and the software that utilize them are usually based on constant travel times between the relevant locations and aim at minimizing total travel time or distance. However, constant travel time assumption is far from reality since the traffic conditions vary by time of day, day of the week, even week of the season or year. Thus, distance/time based optimization does not exactly reflect the real fuel consumptions, hence the actual costs; neither can they be used to accurately account for the GHG emissions. This project fills this gap in the literature and develops efficient algorithms to determine the fuel consumption/GHG emission minimizing path between two nodes on a time-varying transportation networks. Since these algorithms are able to solve minimum weight (cost) path problem on any time-varying network they can be used to solve similar problems in computer, telecommunication, energy distribution networks as well.

Shortest and fastest paths on time-dependent networks can be easily found using existing algorithms from the literature. However, these algorithms are unable to determine the least fuel consuming/ GHG emitting path (“greenest” path) between two nodes. To the best of our knowledge, no such algorithm exists in the literature to efficiently solve this problem. The project presents optimal and effective heuristic methods to determine the greenest path from an origin to a destination on a real road network with time-varying speeds. We first performed an extensive literature review of the shortest path problem and also investigated fuel consumption models to determine the model to be adopted in the experimental studies. Then, we developed the Enhanced Greenest Path Algorithm (EGPA) to find the most fuel consuming/GHG emitting path on time-dependent road networks and tested its performance on different network topologies and with different parameter settings. In addition, a novel approach was presented to efficiently create speed profiles on real road networks and the effectiveness of EGPA was investigated on Washington, DC and Istanbul road networks. Finally, we developed an Ant Colony Optimization algorithm to solve the Time-Dependent Emission Vehicle Routing Problem where the vehicle routes and the corresponding customer-to-customer travel paths were determined simultaneously and examined potential savings and sustainability benefits. 

Ship2Rail: An integrated Service Platform for the Sea-Rail Multimodal Transport Service Providers Based on Revenue Management.

Ship2Rail project is addressing a design and development of a seamless service platform for joint reservation as well as co-sharing in sea-rail multimodal transport in order to create integrated and unified communication opportunities for transport actors.

This project will research, prototype and evaluate an extendable architecture and framework to support quick, safe and easy reservation system among the transport actors in sea-rail multimodal transport. Ship2Rail platform will support multiple heterogeneous execution platforms across different operator domains and over different countries. In this platform, another focus is to define and develop real-time planning algorithms in order to justify the changes in transport planning, adopt different price strategies for transport service providers and also seek the opportunities of co-sharing for transport users in order to reduce empty transport.

For transport users, transport service providers and transport operators, the Ship2Rail project will turn today’s confusing heterogeneity into an easily manageable and rich service environment by exploiting synergies and fostering joint service adoption. The Ship2Rail approach will broaden business opportunities in the sea-rail multimodal transport sector and strengthening the horizontal collaboration among sea-rail multimodal transport providers as well as transport users.

Costs and Benefits of Vehicle Electrification in Turkey

This project aims to assess the costs and benefits of vehicle electrification in Turkey, by analyzing possible future scenarios regarding market penetration rates, different vehicle powertrain options (hybrid electric vehicles, plug-in hybrid electric vehicles and battery electric vehicles), electricity generation and distribution options, as well as transportation behavior and consumer choice.

The success of vehicle electrification will depend on how much the customers will adopt them. There are still some barriers to widespread adoption of these vehicles across the world. Many countries are trying to overcome these barriers with incentives and/or regulations. A similar path might be followed for Turkey, however clever and useful policies will require understanding the costs and benefits clearly.

Benefits of vehicle electrification will strongly depend on vehicle types, driving profiles, driving distances, climate and most importantly the source of electricity generation. These issues can show differences across different regions or cities. Previous literature shows that there might be cases where electric vehicles (EV) don’t prove to be economical in the short term, or environmentally friendly unless electricity generation becomes cleaner.

In addition, in the event that EVs start to penetrate into market, necessary charging infrastructure should be available. Charging infrastructure manifests itself as charging stations and any necessary developments required within the electricity grid. Charging stations can be residential, workplace or public. Either option will have advantages and disadvantages that can manifest themselves in economic, environmental or social aspects. As an example, charging times might change the cost of electricity, and might either increase or decrease CO2 emissions. Workplace or public stations might help people to charge their vehicles while they are at work, however this might require careful infrastructure and regulations planning.

Determining the costs and benefits of vehicle electrification requires understanding all of these issues and their complex interactions with each other. A true understanding of all these issues requires a very comprehensive analysis, involving data collection, socio-economic and techno-economic analyses, physics and/or data based modeling of vehicle energy consumption, energy generation and distribution, policy analysis and policy design, and decision making.

This project is aimed to be an exploratory study, whose main goal is to learn how all the above-mentioned issues should be addressed for Turkey (or, a particular zone in Turkey such as the city of Istanbul). It aims to create a framework for future detailed studies that have high potential of getting academic, public and industrial attention. The project aims to identify currently available data sources, determine what other information would be needed, determine modeling needs and construct models where possible, discuss potential analysis methods and develop a road map for a more comprehensive collaborative multidisciplinary analysis. We also aim to come up with an initial assessment of costs and benefits by using available data and making use of similar information from other countries and models where necessary. We intend to use these preliminary results as a basis for further development of analysis methods. These results might also guide policy makers in the related matter. Finally, with this project, we intend to initiate a collaborative research effort that would make it possible for Sabancı University to have a pioneer role in clean transportation in Turkey.

Robust Risk-Averse Stochastic Programming Models for Multicriteria Decision Making Problems

For many decision-making problems under uncertainty, it is essential to take into account decision makers' risk preferences and specify the preferences based on multiple stochastic performance measures/criteria. Such decision-making problems arise in a wide range of areas, including humanitarian logistics. In the scope of this project, we intend to study various risk-averse optimization models for multicriteria decision making problems and develop associated effective solution methods. In such problems, finding best decisions requires specifying preference relations among vector-valued random variables, where each dimension of a vector corresponds to a decision criterion. In this respect, we consider the multivariate preference relations primarily based on the risk measure called “conditional value-at-risk - (CVaR)''; such relations provide a higher flexibility to express decision makers' risk preferences and allow us to represent a wider range of views. We apply the proposed methods to a stochastic network design problem in a pre-disaster humanitarian logistics context and conduct a computational study concerning the threat of hurricanes in the Southeastern part of the United States.

Building Resilient Economic Agglomerations on Transportation and Health Effects: Urban Form, Location Choice and Transport Solutions for High Air Quality and Low-carbon Cities

We study (i) the interactions between urban form, energy use and emissions by households and firms, and (ii) how the design of policies can contribute to social welfare and compliance with local air quality standards, while speeding up the transition towards prosperous low-carbon cities. A central theme throughout this project is the role of spatial structure in defining sustainable and socially inclusive cities. In European economies increasing city size and urban density tend to reduce households’ average energy consumption. But increasing population density also tends to reduce local air quality. We analyse and contribute the management of this trade-off, by developing (i) surveys to study (environmental) preferences and (location) behavior of firms and households in four different European cities – Amsterdam, Istanbul, Gothenburg, Barcelona; (ii) a spatial general equilibrium model for policy simulations in cities across Europe. In doing so, we take into account that firms and households sort themselves across space, also in response to environmental and transport policies. Our approach is bottom-up and demand driven by involving city governance structures from the outset. We aim to support policy makers in their design of integrated local urban energy and transport systems, assessing the combined effect on economic welfare.

New Models and Efficient Solution Methods for the Electric Vehicle Routing Problem

Road movements have significant economic, environmental, and societal impacts. Transportation has hazardous and threatening impacts on the environment such as the consumption of natural resources, land use, toxic effects on ecosystems and humans, noise, and the effect induced by accidents and Greenhouse Gas (GHG) emissions. Among these, GHG, especially CO2 and NOx emissions as well as the particles are the most concerning since they have direct consequences on human health, such as pollution, and indirect ones, such as the depletion of the ozone layer and global warming. There is an increasing demand in the developed countries and modern societies for reducing these negative impacts and achieving emission-free city centers in particular, and governments have started taking new measures against the vehicles running with fossil fuels. Consequently, the interest has been gradually growing for alternative fuel vehicles such as those moving with solar energy, biodiesel fuels, LNG, and CNG as well as electric vehicles. In parallel with these developments, the research addressing the utilization of electric vehicles in logistics operations has recently gained momentum. Electric vehicles have significant advantages in passenger and freight transport with their zero-emission engines and low energy costs. On the other hand, their high cost of ownership, limited driving ranges, and long recharge durations pose important challenges. These challenges can be prevailed in the logistics activities by planning the electric vehicles operations more effectively. Hence, the route planning of these vehicles has attracted the attention of the researchers and the associated Vehicle Routing Problem (VRP) has been introduced to the literature as the Electric VRP (EVRP). 

The Capacitated VRP (CVRP) is one of the most intensively studied problems in the Operations Research (OR) literature. The problem involves a depot and a set of customers from that depot. The customer demands are satisfied by a homogeneous fleet of vehicles with the same capacity. All vehicles depart from and return to the depot. Each customer is visited exactly once. The objective of the problem is to minimize the total travel cost. VRP with Time Windows (VRPTW) is a variant of CVRP which requires all customers to be served within their pre-determined time intervals. Furthermore, the primary objective may be the minimization of the number of vehicles needed. The most basic forms of both of these problems are very difficult to solve (they are NP-hard). In this project, we will address the VRPTW where the fleet consists of electric vehicles. This problem is referred to as Electric VRPTW (EVRPTW). 

Different from VRPTW, the electric vehicles in the EVRPTW may need to visit stations to have their batteries recharged in order to continue their route. This brings additional complexity to the problem due to the scarcity of the recharge stations and long recharge durations. Recharging may take place at any battery level and the battery may be charged partially or in full. In addition, different equipment may allow the recharging of the same energy amount during different durations. On the other hand, the fast rechargers with high power are more expensive; hence, increasing the unit energy cost. In this project, we will address both single and multiple recharger cases. In addition, we will investigate the cases with linear and non-linear charge duration-energy functions and mixed fleet. With mixed fleet we refer to both heterogeneous fleet of diesel vehicles and electric logistics vehicles with different cargo capacities and driving ranges. We will first formulate different mixed integer programming models of the problems and perform a comparative analysis. Then, we will use these models to develop matheuristics and test their performances. Matheuristics are optimization methods that efficiently combine mathematical programming techniques and (meta)heuristics. Within this context, we plan to utilize Adaptive Large Neighborhood Search method which has been successfully employed for solving many VRPs. We will also devise exact methods to solve the EVRPTW variants and use column generation and branch-and-price techniques to develop effective algorithms.

Efficient Approaches to Find Cost-effective and Visually Attractive Solutions for Multi-Period Vehicle Routing and Scheduling Problems

Traditionally, problems considered in the vehicle routing and scheduling literature usually aim to identify a least cost set of vehicle routes. However, such a routing plan may not always be desirable due to the operational difficulties faced by practitioners especially when the routes are geographically dispersed and intersecting one another. Instead, routing plans with reasonably low –but not necessarily the lowest—costs, and at the same time, with “visually attractive” routes may be easier to adopt and put into practice both by planners and by drivers. Hence, in order to ensure operational convenience for planners and drivers, it may be more convenient to partition the service region into several non-overlapping compact areas, and serve the demand points within the same area by the same vehicle. Moreover, such a routing plan tends to be more robust compared to a plan constructed solely by minimizing cost, because it facilitates responding to unexpected disruptions that may be encountered during the operational stage quickly without causing a significant increase in cost. Even though planners employ vehicle routing and scheduling algorithms (that can compute optimal or near-optimal solutions with respect to traditional objective functions) in order to solve the problems faced in practice, they usually make manual changes in a way to obtain routes that are more compact and do not overlap with each other. This clearly demonstrates the practical need for methods that can produce cost-effective and visually attractive solutions for vehicle routing and scheduling problems.

In this project, mathematical formulations and optimization-based efficient solution techniques will be developed and tested for multi-period vehicle routing and scheduling problems with visual attractiveness and workload balance considerations. To the best of our knowledge, this will be the first study focusing on vehicle routing problems involving the abovementioned real-life aspects within its framework. Proposed formulations and solution methods will likely provide guidance on modeling and solving various rich vehicle routing problems with similar characteristics encountered in practice. Furthermore, the project has the potential to make important methodological contributions to the literature since the problem class that will be considered in this project generalizes different vehicle routing problem variants.

The purpose of this project is to develop efficient methods to solve a class of real-life problems for which there is no systematic solution approach, and to investigate the practical benefits of our proposed methods via computational experiments. We believe that a large audience of decision-makers may benefit from the results obtained within the scope of this project, which aims to facilitate construction of routing plans that are cost-effective, balanced in terms of workload, and easy-to-implement for a problem class that is frequently encountered in distribution –one of the most expensive components of the supply chain— of a product/service.

MeHUB: Integrating a Connected Micromobility Infrastructure to the Existing Public Transport

MeHUB project aims to help organize public space in the urban environment, lower operation costs for micromobility operators and create a better MaaS experience for citizens with the connected and universal micromobility charging infrastructure solution.

MeHUB will:

- Integrate last-kilometer solutions with existing public transportation HUBs.

- Measure scalable benefits of the infrastructure such as, Development of mathematical models and optimization algorithms to determine the optimal locations and sizes for electric HUBs, Change of traffic congestion/carbon emission/public transportation usage by making it accessible to more people.

- Ensure that these vehicles are supported by smart and connected locking against vandalism.

- Minimize operation need, Hence support lowering and measuring effects on carbon emissions.

- Create alternative solutions for efficient and electric personal mobility post Covid-19.