Both Mars One and NASA have suggested sending uncrewed supply ships to Mars as part of a manned mission or colonization. These payloads can take all the time they like, not being constrained by human radiation exposure time limits. Instead, what you want to minimize is the total $\Delta v$ needed to reach Mars.
I can think of two possibilities that have already been used:
1) MOM-style multiple burns at perigee, culminating in trans-Martian injection.
2) Lunar slingshots, perhaps in multiples.
It's possible these two approaches can be combined. There might also be other possibilities, such as starting with a low-energy transfer to EML-2, which is apparently the lowest $\Delta v$ destination in local space.
How do you minimize $\Delta v$ to get to Mars?
EDIT: As Mark Adler points out, it may very well be more efficient to use electric propulsion, which would require higher $\Delta v$ but would result in lower launch mass. However, to avoid making this question too broad, let's assume you're using a high-thrust motor. Perhaps later I'll figure out how to properly ask the underlying question, "What's the cheapest way to send 1000 kg to Mars?"