The term "binary discrete optimization" is almost never used. The terms "combinatorial optimization", "discrete optimization" and "integer programming" are much more common. For more on this, see:
BLP, MPEC and EPEC are usually regarded as nonlinear optimisation problems (i.e., problems with continuous variables and nonlinear constraints). They are however NP-hard in general, since the complementarity constraints are non-convex.
Here is also a nice paper related to a particle swarm optimization algorithm, defined "cat swarm optimization", for binary combinatorial (or discrete) optimization:
When it comes to Binary Optimization, then we should take into account that the decision variable can take just two descrete values, namely "zero" and "one". For example in Traveling Salesman Problem the decision variable (xij) takes one when the salesman travels from city i to city j and otherwise zero.
You can see the topic with a simple level in http://www.amsterdamoptimization.com/pdf/minlp.pdf. With this document, you can also copy and paste the sample codes in GAMS. Note that you can download GAMS free for describing optimization model, and use many MINLP/MILP solvers in http://www.neos-server.org/neos/solvers/index.html to solve your defined problem. Best regards, Pham Duc Dai.