Dynamic Modeling to Identify Mitigation Strategies for Covid-19 Pandemic

Patrick Jenny; David F Jenny; Hossein Gorji; Markus Arnoldini; Wolf-Dietrich Hardt

with α = k d n i|u (0)/k. The best fit with the data was found fork = 0.1654299244 d −1 (the unit d denotes days) and α = 13.75726902; Fig. 2 shows n d computed by the model (line) compared with the data (circles). One can assume that during the initial 50 days detection only occurred due to arising symptoms; therefore the detection time can be set equal to the incubation time. If one further assumes that the incubation time follows a log-normal distribution [12] with 5.84 ± 2.98 days [8] , one obtains an average detection rate of k d = 0.2158 d −1 . As a consequence one obtains for the initial number of undetected infected persons n i|u (0) = αk/k d = 10.5462. Further, we use relation (7) to express k r in terms of k k . With hospital treatment it is assumed that the mortality of the detected persons (fraction of all detected infected persons who die; including everybody who dies) is M = 0.05. Finally, the value of k k was determined to match the reported number of deaths after day 50, which is 1 440, and one obtains k r = 0.1811 d −1 and k k = 0.0095 d −1 . Table 3 shows all parameter values, which we chose for our base case.