Can anybody summerize the binary quadratic programming's definition,
types of problems it applies, the run time issue, and the pros and
cons of algorithms based on it?  ( 750 words )  Thank you.

---

Clarification of Question by robert170-ga on 01 May 2005 04:02 PDT

with sited sources please

---

Request for Question Clarification by mathtalk-ga on 03 May 2005 19:21 PDT

Hi, robert170-ga:

Like "linear programming", the phrase "binary quadratic programming"
(or quadratic binary programming) really conjures up the formulation
of a problem or mathematical model, and not per se an algorithm.

So there are application areas in which these problem formulations are
useful (just factory orders for economic modeling are an application
area for linear programming problems), and there are various
algorithms that one might choose from to seek an optimal solution.

I wonder if you know more precisely what class of model problems is of
interest here.  A simple (but not trivial) form of unconstrained
binary quadratic program is to maximize an objective function:

  u Q u' = SUM SUM Q(i,j)*u(i)*u(j)
            i   j

where the vector u has 0,1 entries.

regards, mathtalk-ga

---

Clarification of Question by robert170-ga on 03 May 2005 23:14 PDT

Thank you for your reply, mathtalk-ga:

This is only for the basic research of quadratic binary programming. 
I would like to know how good it can work on TSP(travel salesman
problem) type of application.

regards, robert170-ga