Hi, I am using CVX for the first time.

My optimization problem is: given a certain p, I need to find lambda and L such that

lambda^(1-p)*L^§ < 1.

My code is below,

The error hints at the requirement of convexity. Can someone give me a hint about how to implement

the above condition?

You can use gp mode http://cvxr.com/cvx/doc/gp.html. You should not be using sdp mode on this problem. Whether you will still be able to if you add other constraints or an objective function is another matter - the problem would still have to be convex, or convertible thereto, in accordance with CVX’s gp rules.

With the problem as is, rather than using gp mode, you could simply take log of both sides of every inequality, and declare variables log_lambda and log_L instead of lambda and L, and have those represent (to you) the respective logs.

In the future, please copy and paste code, using Preformatted text icon, rather than posting image of code.

Please read the following:

Then (re)read the CVX Users’ Guide.