I am using NLPSolve to maximize the the sum of pij*LN(pij) over i and j subject to linear boundary conditions. I notice the help page for NLPSolve says it finds a global max if the objective function is convex. I expect Maple can figure out if x*LN(x) is convex, but I had to define f(x)= xLN(x) if x<>0 and f(0)=0. Does Maple determine whether this function is convex? How? Is there a place for me to declare it convex? Or does the algorithm just al;ways work ( to find a global max) if the objective is convex?