Fix normalization by Monte-Carlo sampled parameters, for short-range interactions
The computation of the probability of infection is incorrect when short-range interactions are included: normalization by several Monte-Carlo sampled parameters is missing in the short-range concentration computation. This is important when getting the probability of infection: the mean (i.e. Monte-Carlo integral) of the diameter-dependent quantities should be computed BEFORE multiplying by other Monte-Carlo sampled parameters (dilution, viral load, breathing rate) and then plugging everything into the probability of infection formula.