why fmincon and patternsearch give me different result for the same problem?

Hi guys, I have a question about the optimizations method. I'm wondering to solve a calculus of variation problem like the Minimal Surface of Revolution using the optimizator, discretizing the problem. I have imposed like boundary condition the point (0,2) and (3,3) where start and finish my catenary function. Moreover, I have imposed the Beltrami identity for each point of the catenary function. Why the fmincon, using SQP solver' give me the result to the problem while the patternsearch diverge and lost the solution? For clarity I have attached the code with fmincon and pattern search. Best Regards, Andrea