You and I know that marketers usually don't want to offer the best product at the lowest price. But the model you're building has never been built before. It doesn't know anything about the world at all. It has to learn what respondents like and what they don't like. If you give it an easy question like whether respondents will like the best product at the lowest price, you allow it to learn.
I like to differentiate attribute combinations that are logically impossible (like square circles, married bachelors and twice two is five) from those that represent unwise marketing (like the best product at a low price or the worst product at a high price). Logically impossible combinations just confuse your respondents and they may give poor answers to what they perceive are poor questions. But while a great product at a low price might be unlikely marketing in the real world, it isn't impossible and respondents do know what to do about it.
For example, let's say it's June 29, 2007 and Steve Jobs just introduced the first iPhone, but he prices it below the price of a flip phone. Would consumers have scratched their heads, unsure how to act? Would they have left the stores, confused over the price tags they saw? I believe they would have bought iPhone and remarked at what a great product it was for the price! Smartphones would have replaced flip phones much more quickly than they did and Apple would be the market share leader instead of Android.
Still, sometimes people do like to add prohibitions to their designs, and you can do this in the software, in the attributes portion of the menu system. You can prohibit the best brands/features from appearing with the lowest prices or the worst brands/features from showing with the highest prices. Some users go through the design manually and remove sets that contain a dominating alternative, though they usually find that doing so doesn't much change their results.