memexp-6.0.exe simple dir2/data.002 1 none 1. 1. 1 40 0 5.0D-4 0
memexp-6.0.exe allin1 1 mem0_no_errors.def dir2d/data.002 1. mem_no_errors.analysis dir2d/data.002_e1 mem0_errors_dif.def dir2d/data.002_e2
The second test case and the final test case in memexp.com involve kinetics simulated as the sum of a sharp (exponential) and a broad phase that overlap in time. They illustrate the benefits of recovering the lifetime distribution by deriving the 'prior' model F from f in two different ways, first by uniform blurring and second by differential blurring [3]. Because the inversion of equation 1 to obtain the distribution of lifetimes is an ill-posed problem, the peak widths are not easily determined. Consequently, it is good to assess the variability in the lifetime distribution that is afforded by the signal-to-noise. This example shows that the changes observed in f upon a reasonable change in F can be used to estimate the range in peak widths consistent with the data. (Note also that the plot size depends on the value of PLOTS specified.)
First, see the lifetime distribution obtained using a uniform model, as called for in the simple run above (IBIGF = 0), plotted in dir2/data.002_e1.out.ps. Then, compare it to the lifetime distribution obtained using differential blurring [3], as called for in mem_errors_dif.def and plotted in dir2d/data.002_e1.out.ps.
For more discussion of differential blurring, see references [1,3].