ABSTRACT
In this work, the optimization-based method is implemented to investigate the effectiveness of lockdown strategies undertaken to contain the COVID-19 during the first two waves in Malaysia. The well-known Susceptible-Infected-Removed (SIR) epidemiological model was fitted to the actual data of infected cases from the official press to closely reflect the observed COVID-19 outbreak in Malaysia. The Genetic Algorithm (GA) and Particle Swarm Optimization (PSO) were implemented to determine the daily transmission rate beta(t) that fits the SIR model to the actual data. The best fitness value of PSO is mostly stable at approximately 37.5 with the best value of 37.41 at a population size of 1000, whilst the best value for GA slowly decreased to the best value of 47.45 at a population size of 1000. In addition, PSO requires a lower number of iterations to reach the optimum fitness value for the same population size as compared to GA, while GA is too far to reach the convergence. As the removal rate (gamma) is a constant value fixed at 0.1, the optimized beta(t) values indicate a high basic reproduction number (average R0 = 1.23) obtained before the Movement Control Order (MCO), followed by a considerable decrease to an average R0 value of 1.23 during the MCO. During the Conditional MCO and Recovery MCO, the basic reproduction number was slightly decreased to an average R0 value less than 1. This is an indication of the success of the government to contain the pandemic during the first two waves as the R0 has been kept below than 1.
ABSTRACT
West Java is in the five line on the list of provinces in Indonesia with the most COVID-19 cases, as Bandung Metropolitan Area (BMA) is the second most densely populated showing the highest number after Jakarta Greater Area. Bandung Metropolitan Area consist of Bandung City, Cimahi City, Bandung Regency, and West Bandung Regency. Then, an intense movement of people created between the connected city and regency. Bandung City became the epicenter of movement BMA, since it is the province capital city, business, and education center. This fact, putting BMA at the highest risk not only for the pandemic but also socioeconomic issues. The spatial time series risk forecasting information is an essential for the decision-maker to develop a day by day policy aimed for combating the COVID-19 pandemic issue. In this study, the pandemic risk is calculated by combining vulnerability, hazard, and geodemography information. Infimap provides the People in Pixels geodemographic data, added not only the exposure of population distribution to COVID-19 but also the ratio of age. Beside those data, the daily distribution of COVID-19 cases, network data, business point, health facility point, residentials area, geodemographic (People in Pixels), and daily COVID-19 Community Mobility Reports is also been used in this study. The daily vulnerability and hazard data created since the first case on March 4th until August 21st. The hazard area is create based on the expected travel area of positive COVID-19 patient. While the vulnerability area is create using Spatial Multi Criteria Analysis (SMCA) of following data: service area of hospital, groceries (local market), and workspace. Further, the time series data of hazard and vulnerability area was inputted to develop the forecasting model based on the machine learning pipeline of Gaussian algorithm. As a result, this study shows the possibility to predict the future risk area of COVID-19 until the next 100 days condition, based on spatial timeseries forecasting model. © 2020 SPIE