Papers by Paola Cappanera
A Decomposition Approach to the Clinical Pathway Deployment for Chronic Outpatients with Comorbidities
AIRO Springer series, 2022
Home-Based and Center-Based Care: From Being Alternatives to Being Synergistic. Optimization Models to Support Flexible Care Delivery
2022 IEEE 16th International Symposium on Medical Information and Communication Technology (ISMICT), May 2, 2022
An efficient management of chronic patients is a key element in the current and future Healthcare... more An efficient management of chronic patients is a key element in the current and future Healthcare context. This paper aims at providing an overview of the PROASSIST 4.0 solution for an healthcare assistance territorial model 4.0. The proposed service relies on the integration between the organizational assets and advanced ICT technologies: a dynamic and adaptive system able to respond to the needs of citizens / patients and to support staff healthcare in the management of an ever-increasing number of patients is proposed. This is achieved by optimizing the scheduling of the territorial assistance based on multiple occurrences and actual patients' health status.
IMA Journal of Management Mathematics, 2013
This paper will demonstrate an orchestration system for 5G infrastructure supporting latency-mini... more This paper will demonstrate an orchestration system for 5G infrastructure supporting latency-minimized and self-adaptive service chaining over geographically distributed edge clouds interconnected through SDN. The demo will show an orchestration system that comprises dynamic virtual function selection and intent-based traffic steering control functionalities to provide optimized service chains in terms of end-to-end latency and adjustable with respect to the context (e.g., service dynamics, network status).
Theory and Practice of Logic Programming, Jul 21, 2023
In answer set programming (ASP), the user can define declaratively a problem and solve it with ef... more In answer set programming (ASP), the user can define declaratively a problem and solve it with efficient solvers; practical applications of ASP are countless and several constraint problems have been successfully solved with ASP. On the other hand, solution time usually grows in a superlinear way (often, exponential) with respect to the size of the instance, which is impractical for large instances. A widely used approach is to split the optimization problem into subproblems (SPs) that are solved in sequence, some committing to the values assigned by others, and reconstructing a valid assignment for the whole problem by juxtaposing the solutions of the single SPs. On the one hand, this approach is much faster due to the superlinear behavior; on the other hand, it does not provide any guarantee of optimality: committing to the assignment of one SP can rule out the optimal solution from the search space. In other research areas, logic-Based Benders decomposition (LBBD) proved effective; in LBBD, the problem is decomposed into a master problem (MP) and one or several SPs. The solution of the MP is passed to the SPs that can possibly fail. In case of failure, a no-good is returned to the MP that is solved again with the addition of the new constraint. The solution process is iterated until a valid solution is obtained for all the SPs or the MP is proven infeasible. The obtained solution is provably optimal under very mild conditions. In this paper, we apply for the first time LBBD to ASP, exploiting an application in health care as case study. Experimental results show the effectiveness of the approach. We believe that the availability of LBBD can further increase the practical applicability of ASP technologies.
Tenant-Side Management of Service Function Chaining: Architecture, Implementation and Experiment on a Future Internet Testbed
2019 IEEE Conference on Network Softwarization (NetSoft)
The adoption of Network Function Virtualization and Software Defined Networking technologies allo... more The adoption of Network Function Virtualization and Software Defined Networking technologies allow network infrastructure operator flexibly orchestrating resources to provide tenants with their own virtual network. However, access to computing and network resource management APIs is typically allowed only within the infrastructure domain and rarely disclosed to tenants for security and performance reasons. This may severely limit tenants capability in coping with demands of application-tailored network services, including Service Function Chaining (SFC). While the literature extensively addressed the challenges of SFC in the infrastructure domain, tenant-side SFC management is quite unexplored yet, although discussed also by a standardization group. This work proposes an SFC platform (called SFCLola) providing tenants with a latency-aware SFC management while minimizing support required from infrastructure operators. The platform encompasses two main levels: an end-to-end chain management level featuring a VNF selection algorithm and a forwarding mechanism that can be programmed and enforced within the tenant network of VMs without requiring access to the switches at the network infrastructure data plane. SFCLola has been implemented as software prototype and experimentally evaluated on a multi-DC infrastructure provided by the 5GINFIRE project.
Computer Networks, 2022
Network Function Virtualization (NFV) and Software Defined Networking (SDN) changed radically the... more Network Function Virtualization (NFV) and Software Defined Networking (SDN) changed radically the way 5G networks will be deployed and services will be delivered to vertical applications (i.e., through dynamic chaining of virtualized functions deployed in distributed clouds to best address latency requirements). In this work, we present a service chaining orchestration system, namely LASH-5G, running on top of an experimental setup that reproduces a typical 5G network deployment with virtualized functions in geographically distributed edge clouds. LASH-5G is built upon a joint integration effort among different orchestration solutions and cloud deployments and aims at providing latency-aware, adaptive and reliable service chaining orchestration across clouds and network resource domains interconnected through SDN. In this paper, we provide details on how this orchestration system has been deployed and it is operated on top of the experimentation infrastructure provided within the Fed4FIRE+ facility and we present performance results assessing the effectiveness of the proposed orchestration approach.
Profit-aware placement of multi-flavoured VNF chains
2021 IEEE 10th International Conference on Cloud Networking (CloudNet), 2021
Network Function Virtualization (NFV) is a promising approach for network operators to cope with ... more Network Function Virtualization (NFV) is a promising approach for network operators to cope with the increasing demand for network services in a flexible and cost-efficient way. How to place Virtualized Network Function (VNF) chains across the network infrastructure to achieve providers’ goals is a relevant research problem. Several emerging aspects, such as possible resource shortage at edge locations and the demand for accelerated infrastructural resources for high-performance deployments, make this problem even more challenging. In such cases, downgrading a service request to an alternative flavour (with less stringent resource requirements and/or fewer offered features) might help increasing the acceptance rate and, to a certain extent, the network service provider’s profit. In this work we formalize the problem of placing network services specified as multi-flavoured VNF chains and present an Integer Linear Programming (ILP) approach for optimally solving it. Simulation results demonstrate the feasibility and potential benefit of the proposed approach, both in online and offline placement scenarios, with an improvement in profit of up to 16% and 18%, respectively, with respect to the case where requests are specified in a single flavour.
Pattern Generation Policies to Cope with Robustness in Home Care
We consider the Robust Home Care problem, where caregiver-to-patient assignment, scheduling of pa... more We consider the Robust Home Care problem, where caregiver-to-patient assignment, scheduling of patient requests and caregiver routing must be taken jointly in a given planning horizon, and patient demand is subject to uncertainty. We propose four alternative policies to fix scheduling decisions and experiment their impact when used as a building block of a decomposition approach. Preliminary experiments on large size instances show that such policies allow to efficiently compute robust solutions of good quality in terms of balancing caregivers’ workload and in terms of number of satisfied uncertain requests.
Springer Proceedings in Mathematics & Statistics, 2016
The use of general descriptive names, registered names, trademarks, service marks, etc. in this p... more The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use.
Decomposition approaches for scheduling chronic outpatients’ clinical pathways in Answer Set Programming
Journal of Logic and Computation
Chronic patients suffering from non-communicable diseases are often enrolled into a diagnostic an... more Chronic patients suffering from non-communicable diseases are often enrolled into a diagnostic and therapeutic care program featuring a personalized care plan. Healthcare is mostly provided at the patient’s home, but those examinations and treatments that must be delivered at the hospital have to be explicitly booked. Booking is not trivial due to, on the one hand, the several time constraints that become particularly tight in the case of comorbidity, on the other hand, the limited availability of both staff and equipment at the hospital care units. This suggests that the scheduling of the clinical pathways for enrolled outpatients should be managed in a centralized manner, taking advantage of the fact that demand for services is known well in advance. The aim is to serve as many requests as possible (unattended requests are supplied by contracted private health facilities) in a timely manner, taking patients priority into account. Booking involves setting a date and a time for each...
Rush order containment of critical drugs in ICUs
PLOS ONE
The recent SARS CoV-02 pandemic has put enormous pressure on intensive care staff, making it impe... more The recent SARS CoV-02 pandemic has put enormous pressure on intensive care staff, making it imperative to relieve them of repetitive tasks with little added value such as drug replenishment. We propose a decision support system based on a hybrid policy to manage the inventory of critical drugs with low and intermittent demand at an Intensive Care Unit (ICU). Demand forecasting is at the heart of any inventory policy. We claim that in the ICU setting drug demand patterns must be therapy based. Heterogeneous data have been collected during an on site study, and information have been extracted to provide a faithful abstract representation of the ward as a system, as well as the potential evolutions of ICU patients clinical conditions. Together with medical guidelines, this provides the foundation of a therapy based demand forecasting tool. This study integrates schedule optimization and demand forecasting, and exploits simulation for evaluation purpose in the long run. At the beginnin...
Metodi e modelli di supporto alle decisioni inerenti la logistica del farmaco
In ambito sanitario, l’utilizzo di metodi quantitativi sembra ormai imprescindibile per risponder... more In ambito sanitario, l’utilizzo di metodi quantitativi sembra ormai imprescindibile per rispondere alla crescente necessità di offrire servizi di assistenza e di cura che sempre più siano centrati sul paziente, ma che al tempo stesso siano anche efficienti e consentano quindi una gestione ottimale delle risorse disponibili, spesso scarse. Questo è il contesto in cui si inserisce lo studio di modelli e metodi di supporto alle decisioni relative alla logistica del farmaco, oggetto di questo studio. In particolare, lo studio ha l’obiettivo di mostrare come l’impiego integrato di tecniche quantitative quali il process mining, l’apprendimento automatico, l’ottimizzazione e la simulazione possano coadiuvare nello sviluppo di politiche di riordino dei farmaci che siano sicure, efficaci ed efficienti
A Two-Phase Approach to the Emergency Department Physician Rostering Problem
Springer Proceedings in Mathematics & Statistics, 2020
In this study, we address the physician rostering problem occurring in an Emergency Department of... more In this study, we address the physician rostering problem occurring in an Emergency Department of an Italian pediatric hospital. Motivated by the paramount importance that workload balance has in this setting, we propose a tailored two-phase approach and we present two optimization models on which the proposed approach is based. In the first phase, we assign all the weekend (and holidays) shifts to physicians in a medium-term planning horizon pursuing a fair distribution of weekend and night shifts among the physicians, whereas in the second phase, we assign all the weekday shifts to physicians in short-term planning horizons so that each physician works almost the same number of morning and afternoon shifts. We present preliminary results of an ongoing research whose ultimate goal is to develop a decision support system to facilitate the creation of physicians’ rosters.
Temporal constraints and device management for the Skill VRP: Mathematical model and lower bounding techniques
Computers & Operations Research, 2020
Abstract We study a generalization of the Skill VRP that incorporates time windows aspects, prece... more Abstract We study a generalization of the Skill VRP that incorporates time windows aspects, precedence and synchronization constraints. Specifically, we are given a logistic network where nodes correspond to customers, and where each customer requires a set of (partially ordered) operations. A set of technicians is available to perform such operations, and each technician is qualified to execute only a subset of them, depending on his skill. By referring to a specific context such as Health Care, customers are patients while technicians are caregivers. In a Field Service context, instead, customers are usually referred to as clients while technicians as field technicians. The innovative aspect is that some operations may require a special device, which must be transported at the customer site and must be present at the customer location together with a technician qualified to use it. Given technician dependent traveling costs, we address the problem of defining the tours for the technicians and for the special device, while respecting the skill compatibility between customers and technicians, and the time windows, precedence and synchronization constraints. We propose a Mixed Integer Linear Programming (MILP) model for the generalized Skill VRP, and present some lower bounding techniques based on the proposed formulation. Preliminary computational experiments show that some lower bounding techniques may rapidly produce good lower bounds, thanks to quite effective valid inequalities. The returned percentage optimality gaps, estimated also thanks to a simple matheuristic, are in fact quite small for several scenarios of medium to large size, by encouraging the use of the proposed lower bounding techniques both as building blocks for designing exact approaches, and also as valuable tools to evaluate the efficacy of more sophisticated heuristic approaches to the problem.
SECOND INTERNATIONAL CONFERENCE ON MATERIAL SCIENCE, SMART STRUCTURES AND APPLICATIONS: ICMSS-2019, 2019
A digital model for the performance prediction of gas turbine performance is presented. The gas t... more A digital model for the performance prediction of gas turbine performance is presented. The gas turbine unit is installed in an industrial site and is equipped with remote monitoring for the main operating parameters. These include temperatures and pressures in key points, power, fuel consumption, and environmental parameters. Different load conditions are investigated, and monitoring includes signals for damage prevention (bearings vibration, lubricant temperatures). An inverse data processing allows to calculate the main performance indexes (system and cycle efficiency; compressor and turbine isentropic efficiency). Different models were evaluated, focusing on ARMAand a Support Vector Machine model for prediction of performance and unexpected events. Possibly leading to alarm/damage conditions. USEFULNESS OF DIGITAL TWINS FOR POWER PLANTS Digital Twins (DTs) are models tuned on experimental data, which should be able to predict accurately the performance of an actual plant [1, 2]. The usefulness of DTs is evident for many purposes, from technical to economical: DTs can be used to drive the actual plants to optimized operating conditions (based on one or more objective functions), or profitfully employed to train personnel, and for predictive maintenance increasing the plant reliability and decreasing O&M costs. Manufacturers are able to provide digital solutions for many new models of their relevant equipment; however, there is a large number of cases of industrial operators who have bought power equipment (often, CHP solutions coupled to important production processes, imposing their specific constraints). These customers need technical/professional assistance for developing a DT model directly from a series of operation data. The model can be built using modern machine learning approaches: it must be validated over different operating
IEEE Transactions on Neural Networks and Learning Systems, 2018
The process of manually labeling instances, essential to a supervised classifier, can be expensiv... more The process of manually labeling instances, essential to a supervised classifier, can be expensive and time-consuming. In such a scenario the semisupervised approach, which makes the use of unlabeled patterns when building the decision function, is a more appealing choice. Indeed, large amounts of unlabeled samples often can be easily obtained. Many optimization techniques have been developed in the last decade to include the unlabeled patterns in the support vector machines formulation. Two broad strategies are followed: continuous and combinatorial. The approach presented in this paper belongs to the latter family and is especially suitable when a fair estimation of the proportion of positive and negative samples is available. Our method is very simple and requires a very light parameter selection. Several medium-and large-scale experiments on both artificial and realworld data sets have been carried out proving the effectiveness and the efficiency of the proposed algorithm.
Stakeholder involvement in drug inventory policies
Operations Research for Health Care, 2019
Abstract This paper experimentally investigates the relationships among three major stakeholders ... more Abstract This paper experimentally investigates the relationships among three major stakeholders that are involved in drug inventory management at Intensive Care Units (ICUs), namely: i) nurses, who in person manage drug orders and carry out storage operations, ii) clinicians, who choose the therapy and shape demand, and iii) the hospital management, who is in charge of the economic sustainability of the hospital. As a case study, we consider the ICU ward of a major Italian public hospital and we focus on antibiotics. We exploit a previously developed Mixed Integer Linear Programming model which decides, for each drug, when and how much to order, and we improve it by adding different sets of constraints to represent each stakeholders’ point of view. By solving three generalized models, each of which ties the satisfaction of a single stakeholder to different thresholds, we explore the mutual effects of taking explicitly into account different perspectives within the inventory policy. We implemented an instance generator, built on the basis of empirical probability distributions extracted from a large set of observed historical data and representing the decision flow ruling drugs prescription. Extensive experiments have been carried out on a set of realistic instances provided by the generator. Results based on our test case not only provide computational evidence to intuitive relations among stakeholders, but also suggest possible levels of compromise. Improved stakeholder satisfaction would also benefit the patient, the passive stakeholder who is the ultimate subject of the caring process.
Empirical Data Driven Intensive Care Unit Drugs Inventory Policies
Springer Proceedings in Mathematics & Statistics, 2017
This paper proposes a drugs inventory policy at point-of-use level, tailored for the Intensive Ca... more This paper proposes a drugs inventory policy at point-of-use level, tailored for the Intensive Care Unit (ICU) case study and aimed at relieving nurses of the time-wasting task of drugs ordering and refilling. The policy aims at jointly reducing order occurrences and imposing service regularity, while keeping stock value as low as possible. An optimization model is proposed and solved on a one-month period real instance and on a set of realistic ones derived from drugs consumption data collection at the ward. The potentially conflicting priorities of three stakeholders (nurses, administration and clinicians) have been successfully incorporated and their impact on order occurrences and stock value has been discussed. Computational results suggest that it is possible to optimize the time-consuming order process currently adopted at the ICU case study. This study is part of a more comprehensive project in which the optimization block will be integrated with a demand forecasting tool and deployed in a rolling horizon framework.
Uploads
Papers by Paola Cappanera