Article
Transportation Science & Technology
Wanjing Ma, Lin Zeng, Kun An
Summary: Flexible buses provide on-demand services based on real-time passenger demand in a cost-effective manner. This study addresses the dynamic bus-routing problem considering stochastic future passenger demand. A two-stage stochastic programming model is formulated to minimize travel time cost and penalty for rejecting requests. The proposed VSC-ALNS algorithm clusters vehicles and passengers based on vector similarity and generates vehicle routes using adaptive large neighbourhood search algorithm.
TRANSPORTATION RESEARCH PART C-EMERGING TECHNOLOGIES
(2023)
Article
Automation & Control Systems
Daniel S. Campos, Joao B. R. do Val
Summary: The article presents an H-infinity-norm theory for stochastic systems in the CSVIU class. It introduces the concept of H-infinity control with infinite energy disturbance signals to accurately represent persistent perturbations in the environment. The article establishes a refined connection between stability and system power finiteness, and utilizes the relations between H-infinity optimization and differential games to analyze the worst-case stability of CSVIU systems.
INTERNATIONAL JOURNAL OF ROBUST AND NONLINEAR CONTROL
(2023)
Article
Economics
Jie Ma, Qiang Meng, Lin Cheng, Zhiyuan Liu
Summary: This study proposes a method to solve the stochastic ridesharing user equilibrium problem in urban transportation network analysis, and its effectiveness is verified through numerical examples.
TRANSPORTATION RESEARCH PART B-METHODOLOGICAL
(2022)
Article
Engineering, Mechanical
Kailing Song, Michele Bonnin, Fabio L. Traversa, Fabrizio Bonani
Summary: In this study, we analyze the performance of a bistable piezoelectric energy harvester with matched electrical load under random mechanical vibrations. The matching network optimizes the average energy transfer to the load. The system is described by a set of nonlinear stochastic differential equations. We find an approximate solution of the stochastic system using a perturbation method in the weak noise limit and use this solution to optimize the circuit parameters of the matching network. In the strong noise limit, we numerically integrate the state equations to determine the average power absorbed by the load and the power efficiency. Our analysis shows that the application of a properly designed matching network significantly improves the performance, with the power delivered to the load improving by a factor of about 17 compared to a direct connection.
NONLINEAR DYNAMICS
(2023)
Article
Operations Research & Management Science
W. Guo, B. Atasoy, R. R. Negenborn
Summary: Global synchromodal transportation involves integrated planning at a network level to move container shipments between inland terminals located in different continents. One challenge is to match accepted shipments with services in a dynamic and stochastic transport network. A sequential decision process model and a reinforcement learning approach are proposed to address the challenge.
ANNALS OF OPERATIONS RESEARCH
(2022)
Article
Computer Science, Interdisciplinary Applications
Javier Andrade-Garda, Juan Ares-Casal, Marta Hidalgo-Lorenzo, Juan-Alfonso Lara, David Lizcano, Sonia Suarez-Garaboa
Summary: This paper discusses the problem of allocating roles to partners in a collaborative network, analyzes the shortcomings of existing centralized matching schemes, and proposes a new solution to address the RPA problem.
COMPUTERS & INDUSTRIAL ENGINEERING
(2022)
Article
Mathematics
Brahim El Asri, Said Hamadene, Khalid Oufdil
Summary: This study investigates the stochastic control-stopping problem with path-dependent and polynomial growth data using the approach of backward stochastic differential equations (BSDEs). The problem is transformed into a specific reflected BSDE with a stochastic Lipschitz coefficient, for which the existence and uniqueness of the solution are demonstrated. The relationship between the solution and the value function of the control-stopping problem is established, and the optimal strategy is presented. In the Markovian framework, it is proven that the value function is the unique viscosity solution of the associated Hamilton-Jacobi-Bellman equation.
JOURNAL OF DIFFERENTIAL EQUATIONS
(2022)
Article
Mathematics, Applied
Thomas Bittar, Pierre Carpentier, Jean-Philippe Chancelier, Jerome Lonchampt
Summary: The stochastic auxiliary problem principle (APP) algorithm is a general stochastic approximation (SA) scheme that iteratively solves a sequence of auxiliary problems to resolve an original convex optimization problem. In this study, we investigate the case where the iterates lie in a Banach space and consider the computation error of the subgradient of the objective. The measurability of the iterates is proven, and convergence results are extended from the Hilbert space case to the reflexive separable Banach space case. Efficiency estimates for function values are derived by adapting the concept of modified Fej\'er monotonicity to the framework.
SIAM JOURNAL ON OPTIMIZATION
(2022)
Article
Economics
A. M. P. Santos, Kjetil Fagerholt, Gilbert Laporte, C. Guedes Soares
Summary: This paper presents a methodology to solve the Supply Vessel Planning Problem with Stochastic Demands (SVPPSD), which uses a two-stage stochastic programming with recourse algorithm. By accounting for the cost of recourse and exploring a wider solution space, robust schedules with a smaller fleet size can be achieved, leading to significant cost savings.
TRANSPORTATION RESEARCH PART B-METHODOLOGICAL
(2022)
Article
Multidisciplinary Sciences
Evgueni Dinvay, Etienne Memin
Summary: We propose a stochastic Hamiltonian formulation for the water wave problem, incorporating location uncertainty. By restricting the general stochastic fluid motion equations to the free surface, we derive the Hamiltonian structure under a small noise assumption. Additionally, the non-local Dirichlet-Neumann operator is explicitly present in the energy functional, enabling systematic approximations and inference of various simplified wave models with noise.
PROCEEDINGS OF THE ROYAL SOCIETY A-MATHEMATICAL PHYSICAL AND ENGINEERING SCIENCES
(2022)
Article
Engineering, Industrial
Sara Ceschia, Margaretha Gansterer, Simona Mancini, Antonella Meneghetti
Summary: The paper introduces the on-demand warehousing problem from the perspective of platform providers and presents mathematical programming and heuristic methods for solving it. The study finds that pricing flexibility on the demand side does not necessarily result in higher payments to the supply side.
INTERNATIONAL JOURNAL OF PRODUCTION RESEARCH
(2023)
Article
Computer Science, Artificial Intelligence
Andreas Look, Melih Kandemir, Barbara Rakitsch, Jan Peters
Summary: Neural Stochastic Differential Equations (NSDEs) model stochastic processes using neural networks. While NSDEs are accurate in prediction, their uncertainty estimation has not been explored. We find that obtaining well-calibrated uncertainty estimations from NSDEs is computationally prohibitive. To address this, we propose a computationally affordable deterministic scheme for approximating the transition kernel of NSDEs, which improves uncertainty calibration and prediction accuracy.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2023)
Article
Economics
Tal Raviv
Summary: This study focuses on the design of a network of automatic parcel lockers to simplify last-mile delivery of small parcels. The network is based on service points near recipients' homes, where they can conveniently collect their parcels. The research aims to optimize the location and capacity of the service points in order to minimize costs and improve service quality.
TRANSPORTATION RESEARCH PART E-LOGISTICS AND TRANSPORTATION REVIEW
(2023)
Article
Statistics & Probability
Ignacio Vidal
Summary: This paper provides a solution to a new approach of Montmort's matching problem from a frequentist point of view. The bivariate probability mass function of the number of matches and the number of unforced errors is derived, and the marginal distribution of the number of unforced errors is expressed using Stirling numbers of the second kind. The distribution of unforced errors is found to be equivalent to that of forced errors.
Article
Computer Science, Interdisciplinary Applications
Feng Zhang, Liwei Zhong
Summary: This paper examines three-sided matching problems with mixed preferences, discussing two different types of preference matching problems and presenting concepts and algorithms for stable matching. Finally, the relationship between these two matching problems is discussed.
JOURNAL OF COMBINATORIAL OPTIMIZATION
(2021)
Article
Multidisciplinary Sciences
Alfredo Braunstein, Alessandro Ingrosso, Anna Paola Muntoni
JOURNAL OF THE ROYAL SOCIETY INTERFACE
(2019)
Article
Physics, Multidisciplinary
Alfredo Braunstein, Anna Paola Muntoni, Andrea Pagnani, Mirko Pieropan
JOURNAL OF PHYSICS A-MATHEMATICAL AND THEORETICAL
(2020)
Article
Physics, Multidisciplinary
Alfredo Braunstein, Giovanni Catania, Luca Dall'Asta
PHYSICAL REVIEW LETTERS
(2019)
Article
Multidisciplinary Sciences
Carlo Baldassi, Fabrizio Pittorino, Riccardo Zecchina
PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES OF THE UNITED STATES OF AMERICA
(2020)
Article
Engineering, Environmental
Ernesto Ortega, Alfredo Braunstein, Alejandro Lage-Castellanos
STOCHASTIC ENVIRONMENTAL RESEARCH AND RISK ASSESSMENT
(2020)
Article
Biochemical Research Methods
Abolfazl Ramezanpour, Alireza Mashaghi
PLOS COMPUTATIONAL BIOLOGY
(2020)
Article
Mechanics
Alfredo Braunstein, Giovanni Catania, Luca Dall'Asta, Anna Paola Muntoni
Summary: This study proposes a novel approach to the inverse Ising problem using the density consistency approximation (DC) to determine model parameters. The comparison with common inference methods shows that DC is one of the most accurate and reliable approaches in inferring couplings and fields. However, there is no single method that is uniformly better than every other one.
JOURNAL OF STATISTICAL MECHANICS-THEORY AND EXPERIMENT
(2021)
Article
Multidisciplinary Sciences
Antoine Baker, Indaco Biazzo, Alfredo Braunstein, Giovanni Catania, Luca Dall'Asta, Alessandro Ingrosso, Florent Krzakala, Fabio Mazza, Marc Mezard, Anna Paola Muntoni, Maria Refinetti, Stefano Sarao Mannelli, Lenka Zdeborova
Summary: Research suggests that probabilistic risk estimation can enhance the performance of digital contact tracing, aiding in mitigating the impact of epidemics.
PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES OF THE UNITED STATES OF AMERICA
(2021)
Article
Physics, Multidisciplinary
Carlo Baldassi, Clarissa Lauditi, Enrico M. Malatesta, Gabriele Perugini, Riccardo Zecchina
Summary: The success of deep learning has revealed the application potential of neural networks across the sciences and opened up fundamental theoretical problems. Statistical physics results suggest that wide flat minima arise as complex extensive structures, formed from the coalescence of minima around high-margin configurations, and have a significant impact on the generalization performance of neural networks.
PHYSICAL REVIEW LETTERS
(2021)
Article
Physics, Fluids & Plasmas
Alfredo Braunstein, Louise Budzynski, Stefano Crotti, Federico Ricci-Tersenghi
Summary: In this study, we investigate a high-dimensional random constrained optimization problem with binary variables subjected to a linear system of equations. We find that different situations occur depending on the random ensemble of linear systems. When each variable is involved in at most two linear constraints, the problem can be partially solved analytically and the optimal solution is obtained upon convergence. We also observe that the system enters a glassy phase with multiple minima in the cost function within a certain range of constraint density. The algorithmic performances are only affected by another phase transition affecting the structure of configurations allowed by the linear constraints. Additionally, we extend our findings to variables belonging to the Galois field of order q and show that increasing q leads to a better optimum, as predicted by the replica-symmetric cavity method.
Article
Optics
A. Ramezanpour
Summary: Reinforcement is applied to the quantum annealing algorithm by adding a time-dependent reinforcement term to the Hamiltonian. The study shows that the reinforced algorithm outperforms the standard quantum annealing algorithm in quantum search and binary perceptron problems.
Article
Physics, Fluids & Plasmas
M. Negri, G. Tiana, R. Zecchina
Summary: The study investigates how to describe the native state of model proteins using physical concepts and sampling algorithms. By efficiently sampling high local entropy states, they demonstrated enhanced stability and folding rate based on simple and general statistical-mechanics arguments.
Article
Mathematics, Interdisciplinary Applications
Indaco Biazzo, Mohsen Ghasemi Nezhadhaghighi, Abolfazl Ramezanpour
Summary: This study investigates the efficiency of movements in cities, particularly focusing on the connection between efficiency and entropy production in transportation processes. Sharing information among selfish drivers enhances predictability in the movement process, while larger cities tend to have smaller efficiencies. Entropy production serves as a good order parameter to distinguish between low- and high-congestion phases, highlighting its role in studying efficiency and predictability in complex systems like cities.
JOURNAL OF PHYSICS-COMPLEXITY
(2021)
Article
Physics, Fluids & Plasmas
Alfredo Braunstein, Thomas Gueudre, Andrea Pagnani, Mirko Pieropan
Summary: The paper introduces a statistical mechanics inspired strategy that addresses the challenge of feature selection from high-dimensional datasets using EP computational scheme. The algorithm, when trained in continuous-weights perceptron, performs well in various conditions and demonstrates robustness and competitiveness.
Article
Physics, Fluids & Plasmas
Anna Paola Muntoni, Rafael Diaz Hernandez Rojas, Alfredo Braunstein, Andrea Pagnani, Isaac Perez Castillo