For the second run, 3,000,000 datasets were simulated for each scenario by building a reference table from a specified set of prior parameter distributions.
Sets of scattered parameter distributions are randomly assigned to the collectors in the solar field.
Using the Latin Hypercube Sampling method (LHS), we generated a sample of set of parameters using distributions of parameter described in table 1 [24], [24].
The one-compartment model is used in contrast to the PBPK model because it employs a relatively small set of parameters whose distributions within the population are generally well characterized.
We found, however, that, for the specific set of parameters and distributions used, the presence of few individuals that become asymptomatic carriers over a considerably long time period (i.e., several months) has no significant influence on the predicted epidemic outcomes.
Once a satisfactory set of parameter values for the distributions is obtained, a Monte Carlo approach is used to simulate pseudo-random values of the variable concerned.
Furthermore, if we want to answer these questions in the most computationally efficient way, we need a procedure for selecting a (somehow) optimal representative set of parameter values from this distribution.
For each scenario and each set of parameter assumptions, the probability distribution of final outbreak sizes according to branching process theory can be calculated (online Technical Appendix) as examples of Lagrangian distributions (28, 29 ).
The theoretical analysis was validated with simulation results and we investigated the behavior of those disciplines for different sets of parameters and distributions for the packet service time and interruption periods.
For each parameter (or a set of parameters), a probability distribution that reflects its prior knowledge/belief is specified and combined with the likelihood function of the data to obtain a posterior distribution of the parameter(s) (e.g., location-specific means of SO2 concentration in our case).
