This extended abstract outlines a novel optimization approach for multi-band linear collaborative spectrum (LCS) sensing based on harmony search heuristics. Besides, the description of a centralized system architecture is also included in this manuscript, which is capable of processing the captured spectrum measurements from several nodes in a LCS fashion and storing the decisions made on the spectrum occupancy. In particular, the value of the coefficients participating in the LSC sensing technique are optimized based on a maximum throughput criteria. The performance rendered by the heuristics utilized in this work are assessed and compared, through extensive simulations, to those of a state-of-the-art layout instead resorting to a genetically-inspired solver for the aforementioned optimization. This work is currently under development under the funded FP7 project CREW (Cognitive Radio Experimentation World).