In building our stochastic discovery, I often encountered a convergence exception.
I am unsure if this is an acceptable solution, but after some testing, I found if I changed the following line to reset the param to lowest or check that the parameter isn't lower than lowest, the convergence exception could be avoided.
In building our stochastic discovery, I often encountered a convergence exception.
I am unsure if this is an acceptable solution, but after some testing, I found if I changed the following line to reset the param to
lowest
or check that the parameter isn't lower thanlowest
, the convergence exception could be avoided.https://github.com/promworkbench/BasicStochasticMiner/blob/8d1485cab435ba51753820149d6e6e01a7fc5cca/src/org/processmining/basicstochasticminer/solver/Solver.java#L89-L91
A change that worked for me was the one of the following:
or