Laura Palagi

Extreme events increasingly challenge the resilience of power grids, complicating postfault restoration under limited resources, incomplete information, and dynamically evolving conditions. This paper proposes a framework for Real-Time adaptive crew scheduling under uncertain and time-varying failure scenarios for Electrical Distribution Networks. To provide a tool for assessing the quality of the proposed approach, we develop an integrated simulation-optimization framework RecSIM-RT, which combines the dynamic evolution of failures and switching operations, with a Mixed Integer Linear Programming model. Uncertainty is addressed through scenario-based simulation of failure arrivals and through path-dependent processing times that account for unknown fault locations. The framework enables re-optimization in response to on-time crew availability and newly detected failures. Computational experiments on realistic instances show that the optimization approach leads to better performance than a benchmark heuristic adopted by electrical operators, with larger gains in high-density and tightly clustered failure scenarios. The real-time optimization mechanism proves to be particularly effective as the scale and frequency of failures increase.

Nowadays, managing Operating Rooms (ORs) in healthcare systems presents a significant challenge, with underutilization of surgical facilities being a key concern. Several decision factors must be considered, including the number of ORs, operating hours, target utilization rates, and the possibility of patient transfers between surgical units (SUs). This study presents a multi-criteria mixed-integer linear optimization model designed for the specific context of Policlinico Umberto I in Rome. The model uses an open scheduling approach to determine the optimal day, time, and OR allocation for weekly surgeries. Using real data from two SUs and four ORs over twelve weeks, we assess the impact of different operating time policies on hospital Key Performance Indicators. Additionally, we introduce randomly generated data to simulate surgery priorities and further validate our model. Results demonstrate the effectiveness of mathematical optimization in improving OR utilization and overall healthcare efficiency.

In this work, we consider minimizing the average of a very large number of smooth and possibly non-convex functions, and we focus on two widely used minibatch frameworks to tackle this optimization problem: incremental gradient (IG) and random reshuffling (RR). We define ease-controlled modifications of the IG/RR schemes, which require a light additional computational effort but can be proved to converge to a stationary point under weak and standard assumptions. In particular, we define two algorithmic schemes in which the IG/RR iteration is controlled by using a watchdog rule and a derivative-free linesearch that activates only sporadically to adjust the stepsize so to guarantee convergence. The two schemes differ in the watchdog rule and the linesearch, which are performed using either a monotonic or a non-monotonic rule. The two schemes allow controlling the updating of the stepsize used in the main IG/RR iteration, avoiding the use of pre-set rules that may drive the stepsize to zero too fast, reducing the effort in designing effective updating rules of the stepsize. We perform computational analysis using different deep neural architectures and a benchmark of varying-size datasets. We compare our implementation with both a full batch gradient method (i.e. L-BFGS) and a fair implementation of IG/RR methods, proving that our algorithms require a similar computational effort compared to the other online algorithms and that the control on the learning rate may allow a faster decrease of the objective function.

This paper addresses the optimal design of wireless networks through the site and power assignment problem. Given a set of candidate transmitters, this problem involves choosing optimal transmitter locations and powers to provide service coverage over a target area. In the modern context of increasing traffic, establishing suitable locations and power emissions for the transmitters in wireless networks is a relevant and challenging task due to heavy radio spectrum congestion. Traditional network design formulations are very ill‐conditioned and suffer from numerical inaccuracies and limited applicability to large‐scale practical scenarios. Our contribution consists of speeding up the solution of the problem under consideration by addressing its drawbacks from a modeling point of view. We propose valid cutting planes and various presolve operations to reduce the problem size and strengthen existing formulations, along with a reduction scheme based on reduced cost fixing to reduce the sources of numerical inaccuracies. Our proposals prove effective, allowing us to achieve optimality on large‐scale instances obtained from a real 4G LTE network in solution times aligning well with planning windows.