You should consider applying to "Operations Research", "Computer science", and "Applied Math" programs (note applied math and not math), all of which can potentially have concentrations in your desired field of study. Make sure the schools have professors publishing in the journals and conferences that interest you. Courses you tend to take in these programs include. Every course below can be used towards a degree in applied math at Cornell (not sure about other places).
Analysis of Algorithms - CS
Simulation - OR
Machine Learning - CS
Combinatorial Optimization - OR
Algorithmic Game Theory - CS
Optimal Learning - OR
Theory of Computing - CS (complexity theory)
Stochastic Dynamic Programming - OR
Networks - CS
Mathematical Programming - code word for analysis with an eye towards optimization and algorithms
Nonlinear Programming - OR
I know that Cornell has a very strong theory component to both their OR and CS programs and you should probably look into them. No matter which program you choose at Cornell, you can bet that you will be able to take a lot of OR,CS and math courses and count most of them towards your PhD requirements (this may not be true at all other universities, you'll have to check). The above list is just a sampling of some of the courses offered at Cornell in OR and CS. Combinatorics is offered in the math department. The difference between all these programs will be in the required courses you have to take, for example OR requires stats, CS requires the theory of programming languages etc., but no one likes every single course they take in their PhD.
Edit: Actually at Cornell, Applied Math has the most flexibility of any program. As long as you have some background in undergrad algebra and analysis (which you can make up during grad school) You can affectively create your own curriculum, with very few requirements. Its very easy to focus on ACO there. Note that many applied math programs are not so flexible, so you really have to look around.