High-Performance Computing and ABMS for High-Resolution COVID-19 Spreading Simulation

Round 1

Reviewer 1 Report

The article is devoted to the approach to modeling and simulating the spread of Covid-19 based on agent-based modeling and modeling. The relevance of the task is dictated by the spread of the COVID-19 epidemic process, as well as the effectiveness of the multi-agent approach to its spread. The aim of the research is not only to support large-scale simulations but also to increase the resolution of simulations. The authors do not assume an underlying network of contacts, and contacts between people responsible for distribution are modeled as a function of geographic distance between people. A switching mechanism is determined that combines models based on radiation and gravity, and switching properties are used at different resolution levels. High-performance computing capabilities were used to simulate millions of simultaneously working agents, each of which displayed human behavior. To carry out such simulations, the authors developed a spreading simulator and tested it by simulating spreading in the two most densely populated Italian regions: Lombardy and Emilia-Romagna.

Despite the satisfactory quality of the article, there are some shortcomings that need to be corrected.

1. The abstract should include not only an introduction and methods but also results.
2. The aim of the research should be defined.
3. A review of existing agent-based models of epidemic process and, in particular, COVID-19, is needed. (e.g. doi: 10.1016/j.procs.2021.12.310 doi: 10.1080/17477778.2021.1965501)
4. It is not clear which data is used for model validation. And has the model verified by real statistics on COVID-19 morbidity?
5. One of the main characteristics of the proposed model is the infection state. But authors did not mention which states they use for the model, and why.
6. It is recommended to discuss the approach of knowledge bases of agents which allow providing intelligent behavior (e.g. doi: 10.1109/ELIT.2019.8892307)
7. Agent-based models' advantage is the possibility of experimental study with factors influencing the epidemic process, but authors concentrate only on forecasting. The results should be interpreted.
8. The Conclusion section should include numerical results of the research.
9. A discussion section should be included and discuss the obtained numerical results and compare the model with existing ones.

In summarizing my comments I recommend that the manuscript be accepted after major revision, including extending the description of the model.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

In this work, the authors present an approach for the modeling and the simulation of the spreading of Covid-19 based on Agent-based Modeling and Simulation (ABMS). This approach permits support for large-scale simulation, as well as increases the simulation resolution. They do not assume an underlying network of contacts, and the person-to-person contacts responsible for the spreading are modeled as a function of the geographical distance among the individuals. In particular, They define a commuting mechanism combining radiation-based and gravity-based models and we exploited the commuting properties at different resolution levels (municipalities and provinces). Finally, they exploit the High-Performance Computing (HPC) facilities to simulate millions of concurrent agents, each mapping the individual’s behavior. To do such simulations, they developed a spreading simulator and validated it through the simulation of the spreading in two of the most populated Italian regions: Lombardy and Emilia-Romagna.

The manuscript is well presented and the results seem correct. Nevertheless, according to the Covid-19 pandemic in Italy, it is judicious if the authors integrate the disease-induced death in their model. Indeed, Italy is one of the countries which recorded several deaths at the beginning of the disease. Thus, it is important to take into account the fact that one agent can die after a virus infection. Also, the authors must integrate the graphs of Daily death persons with different tp values 

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

Although the paper is interesting and promising it requires a lot of rework in order to be published.

The main concern is that there is a lack of technical details about the simulation model and the implementation of the simulator.

In detail:

Section 1 or 2: You must detail what are the improvements or differences over existing models and simulators. I don't see a state-of-the-art review.  

Section 3.2: It is extremely lengthy and unnecessary. Please shorten it to explain only the minimum details required to understand your simulator,

Section 4: More technical details are required about your model and simulator. Architecture, design and so on are missed.

Section 6 is too short and the HPC description did, do not justify this section. If you want to evaluate HPC, you should perform more experiments. For example, executing your simulations with 1,2,4,8 processor and show the obtained speed up.

Minor issues.

Pag.8: Numerate the equations

Pag.8 Why do you use these constants?

Pag.8 In the second equation p_i is described below but it is not in the equation.

Pag. 10. What is the meaning of the employment rate?

Ref 4 is from 2012 not 2021.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Thanks for the authors for considering comments and recommendations. In my opinion, now article can be accepted in present form.

Reviewer 2 Report

This version of the manuscript is well improved. I recommend it for publication.

Reviewer 3 Report

Thank you for your response and extensive revision.