Computer simulation of complex conformational transitions in biomolecules, Such as protein folding, is considered one of the main goals of computational chemistry. A recently developed methodology, bias-exchange metadynamics, was Successfully Used to reversibly fold some small globular proteins. The objective of this work is to further improve this promising technique. This is accomplished by searching for the optimal set of parameters that enable folding a 48 amino acid protein, 1E0G, in the shortest possible time, using a coarse-grained force field UNRES. It is shown that bias-exchange metadynamics, if appropriately optimized, allows finding the folded state of 1E0G significantly faster than normal replica exchange.