I would not put in the prohibitions, no. You have to let the design explore the entire design space in an even and balanced way. It might "feel like" you're wasting tasks by showing unrealistic concepts, but that feeling is deceiving you: what you're really doing is creating a balanced experiment, one that will draw on the power of the experimental design to give you precise estimates of all the utilities.
I would put in prohibitions in cases where logical impossibilities would result without them (e.g. yo'd have combinations like "married bachelors" or "square circles" or "twice two is five" or "motorcycle with sunroof," or "Tylenol tablets, 8.6 fl oz." Those will confuse respondents who will rightly be frustrated by your survey. They are usually best handled using alternative specific designs.