Papers by Bilge Küçük Atasoy
From theORy to application: learning to optimize with Operations Research in an interactive way
arXiv (Cornell University), May 12, 2023
In Inverse Optimization (IO), an expert agent solves an optimization problem parametric in an exo... more In Inverse Optimization (IO), an expert agent solves an optimization problem parametric in an exogenous signal. From a learning perspective, the goal is to learn the expert's cost function given a dataset of signals and corresponding optimal actions. Motivated by the geometry of the IO set of consistent cost vectors, we introduce the "incenter" concept, a new notion akin to circumcenter recently proposed by Besbes et al. (2023). Discussing the geometric and robustness interpretation of the incenter cost vector, we develop corresponding tractable convex reformulations, which are in contrast with the circumcenter, which we show is equivalent to an intractable optimization program. We further propose a novel loss function called Augmented Suboptimality Loss (ASL), a relaxation of the incenter concept for problems with inconsistent data. Exploiting the structure of the ASL, we propose a novel first-order algorithm, which we name Stochastic Approximate Mirror Descent. This algorithm combines stochastic and approximate subgradient evaluations, together with mirror descent update steps, which is provably efficient for the IO problems with discrete feasible sets with high cardinality. We implement the IO approaches developed in this paper as a Python package called InvOpt. Our numerical experiments are reproducible, and the underlying source code is available as examples in the InvOpt package.
Modelling Route Choice in Crowd Evacuation on Passenger Ships
International Journal of Maritime Engineering, 2021
This paper proposes an agent-based simulation model with route choice process to predict the crow... more This paper proposes an agent-based simulation model with route choice process to predict the crowd behaviours and evaluate the evacuation safety on passenger ships. The model focuses on the behaviours of two common types of passengers that are not typically accounted for during most evacuation analyses, namely, passengers who are not familiar with the ship layout and passengers who have family members or friends with them. In the proposed model, a marker concept is introduced to represent critical routing points of the layout and passenger agents make a route choice based on their surroundings and characteristics instead of just following the shortest routes. The simulation model is tested by two small but targeted scenarios and one comprehensive scenario on a ship deck. For ship designers, a more realistic evacuation time is provided to better assess the evacuation performance of a ship, and a heat map of crowd density is presented to identify possible bottleneck areas.
Maritime Economics & Logistics
Container terminal capacity is often limited by (in)efficiency bottlenecks. This paper provides t... more Container terminal capacity is often limited by (in)efficiency bottlenecks. This paper provides the design and proof of concept for the bottleneck mitigation cycle (BMC), consisting of three steps: bottleneck classification, detection and alleviation. While, often, the literature only focuses on alleviation of a single bottleneck and ignores bottleneck detection and interdependencies, this research uses the shifting bottleneck method and thereby considers a variety of possible infrastructural and operational bottlenecks. An empirical approach is adopted to find the cause of the detected bottleneck and to suggest suitable alleviation measures. Application of the BMC to a simulation model of the Fergusson Container Terminal in the Port of Auckland resulted in productivity improvements of 2-6%. To further improve the BMC, future research directions are to improve the empirical approach used for bottleneck alleviation and to apply the BMC in real-time.
Models and algorithms for integrated schedule planning and revenue management
Impact analysis of a exible air transportation system: Clip-Air
Smart Mobility through Personalized Menu Optimization
Transportation Research Board 96th Annual MeetingTransportation Research Board, 2017
A choice-based optimization approach in flexible mobility on demand
Eurosurveillance, 2016
A Federated Platform Enabling a Systematic Collaboration Among Devices, Data and Functions for Smart Mobility
IEEE Transactions on Intelligent Transportation Systems, Apr 1, 2023
Users may download and print one copy of any publication from the public portal for the purpose... more Users may download and print one copy of any publication from the public portal for the purpose of private study or research. You may not further distribute the material or use it for any profit-making activity or commercial gain You may freely distribute the URL identifying the publication in the public portal If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.
IEEE Open Journal of Intelligent Transportation Systems
Her research interests lie at the intersection of optimization and behavioral models with applica... more Her research interests lie at the intersection of optimization and behavioral models with applications to transportation systems. More specifically, she is interested in improving the efficiency, robustness, and sustainability of transportation and logistics by developing adaptive models that incorporate the preferences of decision makers. Example applications include on-demand transportation, intermodal freight transportation, and transportation over water.
Compared to a fleet composed of standard planes... Clip-Air carries on the average 5-10% more pas... more Compared to a fleet composed of standard planes... Clip-Air carries on the average 5-10% more passengers by using 20-30% less capacity, Clip-Air copes better with the insufficient transportation capacity, Uses less aircraft/wing ⇒ less crew, simplified operations. The scheduling decisions are robust to the estimated cost figures of Clip-Air. Sensitivity analysis showed that Clip-Air is always better in the number of transported passengers and generates higher profit in 89% of the instances. A preliminary analysis is carried out on the multi-modality aspect of Clip-Air where the empty capsules are allowed to be repositioned via railways. The results show a potential increase in profit and transportation capacity. Supply ⇒ Flexibility provided by Clip-Air Demand ⇒ integration of appropriate demand models Demand model Simple models (e.g. linear, exp.) fail to represent the reality Integrated model becomes very sensitive to demand model parameters Appropriate models need to be developed 5/ 18 Revealed preferences (RP) data: Booking data from a major European airline Lack of variability Price inelastic demand 6/ 18 maximum price elasticity among economy itineraries is −2.16, and −1.95 for business. maximum time elasticity among economy itineraries is −0.33, and −0.36 for business.
In this paper, the authors present a real-time toll optimization framework where the toll optimiz... more In this paper, the authors present a real-time toll optimization framework where the toll optimization is integrated with a mesoscopic traffic simulator, DynaMIT, so that the tolls are optimized based on predicted traffic conditions. DynaMIT embeds several modules including demand simulation, supply simulation, and online calibration. The toll optimization module is in complete interaction with DynaMIT such that the optimized toll is decided with several iterations between the two rather than a single feedback function. Two main formulations are proposed; the first maximizes revenue and the second considers also the traffic conditions on managed lanes while maximizing revenue. The objective of this paper is to analyze the framework on a simple network in order to show its potential and understand the impact of several factors. Different numerical experiments are presented with different formulations and behavioral assumptions. It is shown that the framework generates consistent resu...
Augmented Lagrangian relaxation-based coordinated approach for global synchromodal transport planning with multiple operators
Transportation research. Part E, Logistics and transportation review, May 1, 2024
Modelling and Optimisation of Ship Energy Systems 2023, Jan 6, 2024
Several measures have been developed to prevent emissions from inland water transportation. Howev... more Several measures have been developed to prevent emissions from inland water transportation. However, it is challenging to weigh all the aspects to identify the pathway that will ultimately result in zero-emission inland shipping. A data-driven virtual representation of the inland shipping system can be used to evaluate zero-emission strategies, effectiveness of policies and technologies, and consequences of their implementation. This multi-level digital twin can realistically represent the system with all relevant components, which needs to be validated using real-world data. Subsequently, future scenarios can be imposed on the digital twin, and the proposed intervention measures can be applied, based on which their efficiency can be assessed together with the inland shipping sector. This study discusses the essential aspects of designing a digital twin for an IWT. Three aspects are considered essential: individual ships, logistics chains, and infrastructure. As these research topics span various scales, ranging from a single vessel to an entire infrastructure network, an agent-based approach is suitable for forming the basis of the digital twin. Consequently, potential interventions can be considered, ranging from the application of new technologies to individual vessels to policy measures implemented for an entire shipping corridor or various bunker infrastructure strategies in the network. Additionally, the impact of the implemented interventions can be evaluated at any desired scale, ranging from the individual ship level and its emissions to the network level and aggregated emissions in an entire area, or the impact on the logistics chain.
arXiv (Cornell University), Nov 26, 2023
The design and pricing of services are two of the most important decisions faced by any intermoda... more The design and pricing of services are two of the most important decisions faced by any intermodal transport operator. The key success factor lies in the ability of meeting the needs of the shippers. Therefore, making full use of the available information about the demand helps to come up with good design and pricing decisions. With this in mind, we propose a Choice-Driven approach, incorporating advanced choice models directly into a Service Network Design and Pricing problem. We evaluate this approach considering three different mode choice models: one deterministic with 4 attributes (cost, time, frequency and accessibility); and two stochastic also accounting for unobserved attributes and shippers' heterogeneity respectively. To reduce the computational time for the stochastic instances, we propose a predetermination heuristic. These models are compared to a benchmark, where shippers are solely cost-minimizers. Results show that the operator's profits can be significantly improved, even with the deterministic version. The two stochastic versions further increase the realized profits, but considering heterogeneity allows a better estimation of the demand.
A real-time synchromodal framework with co-planning for routing of containers and vehicles
Transportation Research Part C: Emerging Technologies, Nov 30, 2023
An assessment methodology for a modular terminal concept for container barging in seaports
Case Studies on Transport Policy, Nov 30, 2023
Zenodo (CERN European Organization for Nuclear Research), Nov 17, 2022
This study examines how to eliminate the inefficiencies in inland waterborne transport (IWT) by r... more This study examines how to eliminate the inefficiencies in inland waterborne transport (IWT) by reducing sailing and waiting times at ports without expensive modifications to terminal infrastructures. A new concept known as the Modular Mobile Terminals (MMTs) is examined to achieve this and its potential feasibility in seaports. In doing this, the present paper develops a time savings optimization model that evaluates regions that would be suitable to be linked based on the time gains that can be derived by using the MMTs. In addition, a cost model is developed to determine the net savings of the actors and the positive business case that can be derived by using the MMTs. Results revealed that regions where vessels have low load factors, would have a positive business case for using the MMTs. The small call sizes transported from the regions can be consolidated at the MMTs and transferred in high volume to dedicated deepsea terminals. This reduces the port time for the barges involved while ensuring that the barges are handled quicker and better. Finally, this system would help enhance the deepsea terminal operations by ensuring guaranteed high cargo volumes at fixed and regular time slots.
Uploads
Papers by Bilge Küçük Atasoy