The Pareto frontier is about tradeoffs. Systems can typically be optimized along one dimension only if there are tradeoffs along others. You can build as cheaply as possible, but you sacrifice beauty. You can maximize how much you work, but you’ll have less time for play. You can maximize an engine’s horsepower, but you will decrease fuel efficiency.
The Pareto frontier is defined by the set of solutions that are equally good. If you want fuel efficiency x, the best possible engine will get you y horse powers. I you have a bad engine, you may still get to fuel efficiency x, but you’ll get less than y horse powers. This means that you’re behind the Pareto frontier. Similarly, an engine with y horse powers but worse than x fuel efficiency will also be behind the Pareto frontier. Of course, there will also be other tradeoffs, such as with how much noise the engine makes, how long it lasts, how cheap it is to make, how long it lasts etc. The Pareto frontier defines how far you can optimize all these parameters given their tradeoffs. How far can you push that envelope? The Pareto frontier isn’t necessarily fixed: Occasionally, an innovation will come along that allows extending the Pareto frontier beyond its current boundary.
The Pareto frontier isn’t the only concept named after polymath Vilfredo Pareto. There’s also the unrelated Pareto principle, also known as the 80/20 principle. It’s a special case of the power law distribution: 80% of wealth belongs to 20% of the population, 80% of complaints are due to 20% of the residents, etc.
Nehaveigur 