How does one get the value of the penality parameter mu at the end of an ipopt run (perhaps maximum number of iterations was reached) and how does one use that value during a warm start that continues the optimization ?