The ongoing epidemic spread of COVID-19 constitutes a global problem with consequences at different social levels (health and mortality, sanitary systems, public resources allocations, economics, job market, etc). An in-depth analysis of the measures implemented by the different Governments to manage the pandemic is useful to highlight their effectiveness and the social consequences of such decisions. Their examination is relevant not only in case of possible re-infection, but also for possible future epidemic risk events in such a globalized world. The COVID-19 emergency and the consequent reactions inspire this project, representing a case study for understanding the causes-effects relationships and for starting to define a generalization to future pandemics risk management. Epidemic emergency management must be faced from multiple points of view (political, economic, social, medical, sanitary, etc) and requires a highly complex analysis of real data, that sometimes (especially in the pandemic initial stages) are poor and collected inconsistently.
With the techniques of modeling, identification, control and optimization, the virus spread and its effects can be studied at different levels, for example studying categories of a single population ¿ grouping people according to some criteria, (the age, the vulnerability to the virus, the possibility of access to treatments ..) or considering the interactions among populations. Each study provides useful information to understand what happened and, possibly, to predict what could happen in similar cases.
This project aims to provide a deep analysis of the current pandemic emergency and a unified framework for risk management through the definition of optimal decision models and procedures for managing and controlling the spread and the effects of the epidemic. Ready-to-use protocols and possible specific interventions on the population will be proposed to reach optimal economic and social health-driven choices.
The COVID-19 spread has shown the great vulnerability of all the nations with respect to unknown and unexpected viruses, especially nowadays in which infected individuals may rapidly move worldwide taking less than most viruses incubation time. Moreover, whenever new unknown viruses spread, any medical action is weakened by the difficulty of understanding important related phenomenon such as how they affect both human and animal population, the diseases they can introduce and their impact on the public health services. The situation is also worsened by the persistent lack of vaccines as well as specific antiviral drugs. These peculiarities highlight the main priority induced by new pandemics: controlling the spread especially during the initial and the most dangerous time by using an efficient combination of different actions. Notable among these actions is social distancing, which plays a fundamental role, but in conjunction to additional different solutions.
This project intends to propose an approach based on the identification of multiple models able to describe different aspects of the pandemic; the aim is to develop combined efficient strategies to contain the spread by using optimization and optimal control techniques. Objective indicators to evaluate the impact of the control actions undertaken and to predict the effect of other strategies are needed; their definition, along with their validation based on the experience acquired in these months of COVID-19, is one important goal of this project.
The main innovations are the following, Fig .5:
---consider interesting aspects of COVID-19 to be suitable modeled; while some of them have already been cited (age factors, vulnerabilities, geographical influences, pollution), others could be detected after a deep analysis of real data available from different regions, even if not intuitively and immediately related to the epidemic spread. This implies large data collection, analysis, modeling, and identification.
---model the possible containment measures; mathematically, they are given by the control actions to be properly included in the model. The governments have imposed social distancing, closed schools, offices and shops, supported smart working whenever possible, required the use of masks and plastic gloves in sheared spaces, limited or even inhibited travels among regions or countries, prohibited occasions for crowding people, like concerts, theater and cinema show, religious ceremonies, and so on. These actions must be translated into mathematical expressions by using suitable data analysis and machine learning techniques.
--- consider suitable integrated analysis of the controlled models; being the diffusion and control of COVID-19 a complex phenomenon, involving not only the medical aspects but also political choices, economical effects, social consequences, job market and enterprises reorganizations, psychological aspects, and so on, the focus of the project will be the integration of all the information obtained studying the COVID-19 from all the different points of view highlighted by the analysis of available data. This integration aspect is both the most interesting and the hardest task, since the relations to be modeled regard aspects of different nature which must be inferred by data analysis techniques;
--- provide objective indicators to evaluate the impact of the applied containment measures and of possible different choices.
--- develop decision support systems by means of operations research models building. These systems intend to help decision makers in charge of epidemic related issues within their area of responsibility. As some of the aforementioned examples, these types of models will be extremely useful in a wide variety of scenarios.
--- In order to better orient mathematical programming performance in terms of their actual problem-solving power, some of the optimization models input coefficients will be obtained by the outputs of the epidemic optimal control models. These proposed integrated approaches represent a promising innovation aspect of this project.
--- With the new opportunities arising from the current expansion of digitalization, planning and management need to be revisited by taking a data-driven perspective. Indeed, extracting knowledge from data helps in reducing uncertainties using predictions and supports the identification of the causes of inefficiencies, disruptions, and anomalies. This project aims to set up a data-driven analytic approach for supporting epidemic crisis management allowing to monitor, control and optimize the necessary policies and resources.
Each epidemic has its own characteristics and generalizations is a process to be carried on with high carefulness; nevertheless, general ready-to-use protocols to be applied before the epidemic turns to a pandemic are useful for a prompt homogeneous response.