In this paper we introduce a novel swarm intelligence approach, based in the artificial bee colony optimization algorithm (ABC), now designed specifically to solve constrained numerical optimization problems: The scout-behavior modified artificial bee colony (SM-ABC) algorithm. In SM-ABC, the behavior of the scout bee is modified as to get the capability to exploit the vicinity of the current best solution (food source). Also, the way to control the tolerance for equality constraints is altered. SM-ABC looks to improve the capabilities of ABC to find good solutions in problems with a high dimensionality and active constraints. The performance of SM-ABC is tested in 13 well-known benchmark problems found in the literature. A comparison is performed between the published results of the original ABC algorithm and SM-ABC. Finally, a performance comparison against algorithms from the state-of-the-art in bio-inspired constrained optimization is shown. The results suggest that SMABC is a promising heuristic to solve numerical constrained optimization problems.