A COMPUTATIONAL MODEL OF EPIDEMIC PROCESS ON A SYNTHESIZED HUMAN INTERACTION NETWORK

• Main Author: Seprianus
• Format: Dissertations
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/86870
• Institution: Institut Teknologi Bandung

In general, compartmental epidemic models assume the population is homogeneous, meaning that each individual has the same chance, per unit time, of interacting with every other individual in the population. As a result, each individual has the same opportunity to be infected or infect other individuals. However, most individuals in a population interact with only a small number of other individuals. This fact indicates that using homogeneous population assumption in epidemiological modeling is no longer relevant. The study in this dissertation aims to develop an epidemic model considering the diversity of human interaction behavior. This model begins by building a network that can describe the diversity of human interaction behavior in a population. This network is called a synthesized human interaction network built using the average degree and a range of interaction intensities. The mechanism for transmitting infection in this network is built by bringing the concept of the common compartment model to an individual scale. This model is built in such a way that a variety of intervention measures can be included without having to add compartments. There are three studies included in the dissertation and are part of the process of developing an epidemic model on a synthesized human interaction network as well as an application of the model that has been developed. The three studies are as follows. The study entitled Modeling the Spread of COVID-19 in Schools and Workplaces: A Computational Approach uses an epidemic model on a synthesized human interaction network to determine the number of individuals who can be accommodated in a room with a certain area and occupancy time. This research builds a synthesized human interaction network based on distances between individuals obtained from random movements of individuals in the room. The epidemic model in this network was developed from the SIR model brought to the individual scale. Several intervention measures such as wearing masks, physical distancing, maintaining hygiene, and reducing outdoor activities are also included without adding compartments. The simulation results show that the number of individuals who can be accommodated in a room when accommodating intervention measures is slightly greater than in the case without intervention as long as the occupancy time is not more than seven hours. The study entitled A Computational Model of Epidemic Process with Three Variants on a Synthesized Human Interaction Network uses an epidemic model on the synthesized human interaction to see how the dynamics of the epidemic process with three variants that emerge randomly with a small probability during the epidemic process. Network models are built using the average degree and a range of interaction intensities. The emergence of new variants during the epidemic is considered a virus mutation. The epidemic model in this study was developed from the SIR model for the individual scale with the addition of two variants. There were no intervention measures accommodated explicitly in this study. However, reducing the average degree and range of interaction intensities can be considered as containment measures. The simulation results show varied and rich dynamics of the epidemic process. The average degree and intensity range of interactions play a large role in determining the size of an epidemic. This research also introduces the infection spread number to measure the severity of an epidemic based on the size and duration of the epidemic. The study entitled A Simple Modeling of the Epidemic Process with Two Vaccine Doses on A Synthesized Human Interaction Network also uses an epidemic model on the synthesized human interaction network to see the impact of two-dose vaccination on the epidemic process. The network model built is similar to the previous research, using the average degree and the range of interaction intensity. The epidemic model on the network is built based on the SIR model for the individual scale. The difference with previous studies is that it includes two-dose vaccination based on the time of vaccine administration and its distribution, namely randomly or based on segment coverage, as an intervention measure. The simulation results show that the two-dose vaccination that is more effective and efficient in curbing the epidemic is vaccination that is carried out earlier and based on segment coverage, namely by targeting hubs in the network.