The tool switching problem with non-uniform tool sizes is considered in thi
s paper for the first time in a dynamic (online) environment. We develop th
ree heuristics, designed for various assumptions with respect to the random
ness of the process; the heuristics are based on the insight and experience
that we obtained from the static problem with non- uniform tool sizes. Our
computational experiments which examined various test problems with differ
ent underlying assumptions indicate that our heuristics are performing well
, that is, they are reasonably close to the static solution and perform muc
h better than a random solution.