Feasibility Problem-UAV communication

I want to know is there any methods to find the number of feasible answers when solving “feasibility problem”

I initially predefined some points (eg: 150 locations) to deploy the UAV, and i have SINR related constrains for every points. when i solve this via CVX, it show only solved or infeasible.

but i want to know how many points are feasible out of predefined 150.

actually need Matlab based answer

Are all the variables integer (or binary)? I doubt there’s a good general way of doing this in CVX, but there are probably ways of adding constraints to exclude certain already found solutions. If your problems are of the correct type, you might be able to use Gurobi or CPLEX solution pools, from MATLAB. but not via CVX - further discussion of those would be off-topic for this forum, as they are not available under CVX.

Anyhow, if there are only 150 points, and presuming you can enumerate them, then why not evaluate the constraints at all 150 points using MATLAB, without an optimizer, and just count how many of the points satisfy the constraints?

I’ll check other solutions too…

variables are binary, as a simplified problem only have 150 points… but the real problem will be more points. further i plan to add some constrains and variables as well.

therefore i need to find a solution for this.

You haven’t told us what kind of constrains your problem has.If linear or convex quadratic, perhaps you can use CPLEX solution pool https://www.ibm.com/developerworks/community/forums/html/topic?id=d8d12e36-3150-4e1c-a956-d707d17f274c&ps=25 . And I believe something similar using Gurobi. Although Gurobi is callable from CVX, I don’t believe the solution pool can be accessed from CVX.


Constrains are SINR of the ground users, flying height, restricted zone.

What matters is what mathematical form they have. Are they linear (affine) in the decision )optimization) variables? Are they convex quadratic or expressible as Second Order Cone constraints? Exponential Cone? (Linear) Semdefinite? if none of the above, then CVX is not applicable at all to the problem.