Balazs Kulcsar

Vincent Szolnoky, Viktor Andersson, Balazs Kulcsar et al.

Most complex machine learning and modelling techniques are prone to over-fitting and may subsequently generalise poorly to future data. Artificial neural networks are no different in this regard and, despite having a level of implicit regularisation when trained with gradient descent, often require the aid of explicit regularisers. We introduce a new framework, Model Gradient Similarity (MGS), that (1) serves as a metric of regularisation, which can be used to monitor neural network training, (2) adds insight into how explicit regularisers, while derived from widely different principles, operate via the same mechanism underneath by increasing MGS, and (3) provides the basis for a new regularisation scheme which exhibits excellent performance, especially in challenging settings such as high levels of label noise or limited sample sizes.

Fangting Zhou, Balazs Kulcsar, Jiaming Wu

Electric mobility faces several challenges, most notably the high cost of infrastructure development and the underutilization of charging stations. The concept of shared charging offers a promising solution. The paper explores sustainable urban logistics through horizontal collaboration between two fleet operators and addresses a scheduling problem for the shared use of charging stations. To tackle this, the study formulates a collaborative scheduling problem as a bi-objective nonlinear integer programming model, in which each company aims to minimize its own costs, creating inherent conflicts that require trade-offs. The Balanced Bounding Box Methods (B3Ms) are introduced in order to efficiently derive the efficient frontier, identifying a reduced set of representative solutions. These methods enhance computational efficiency by selectively disregarding closely positioned and competing solutions, preserving the diversity and representativeness of the solutions over the efficient frontier. To determine the final solution and ensure balanced collaboration, cooperative bargaining methods are applied. Numerical case studies demonstrate the viability and scalability of the developed methods, showing that the B3Ms can significantly reduce computational time while maintaining the integrity of the frontier. These methods, along with cooperative bargaining, provide an effective framework for solving various bi-objective optimization problems, extending beyond the collaborative scheduling problem presented here.

Fangting Zhou, Attila Lischka, Balazs Kulcsar et al.

This paper reviews the current progress in applying machine learning (ML) tools to solve NP-hard combinatorial optimization problems, with a focus on routing problems such as the traveling salesman problem (TSP) and the vehicle routing problem (VRP). Due to the inherent complexity of these problems, exact algorithms often require excessive computational time to find optimal solutions, while heuristics can only provide approximate solutions without guaranteeing optimality. With the recent success of machine learning models, there is a growing trend in proposing and implementing diverse ML techniques to enhance the resolution of these challenging routing problems. We propose a taxonomy categorizing ML-based routing methods into construction-based and improvement-based approaches, highlighting their applicability to various problem characteristics. This review aims to integrate traditional OR methods with state-of-the-art ML techniques, providing a structured framework to guide future research and address emerging VRP variants.

Balazs Varga, Balazs Kulcsar, Morteza Haghir Chehreghani

In this paper, we place deep Q-learning into a control-oriented perspective and study its learning dynamics with well-established techniques from robust control. We formulate an uncertain linear time-invariant model by means of the neural tangent kernel to describe learning. We show the instability of learning and analyze the agent's behavior in frequency-domain. Then, we ensure convergence via robust controllers acting as dynamical rewards in the loss function. We synthesize three controllers: state-feedback gain scheduling H2, dynamic Hinf, and constant gain Hinf controllers. Setting up the learning agent with a control-oriented tuning methodology is more transparent and has well-established literature compared to the heuristics in reinforcement learning. In addition, our approach does not use a target network and randomized replay memory. The role of the target network is overtaken by the control input, which also exploits the temporal dependency of samples (opposed to a randomized memory buffer). Numerical simulations in different OpenAI Gym environments suggest that the Hinf controlled learning performs slightly better than Double deep Q-learning.