Page 208 - Distributed model predictive control for plant-wide systems
P. 208

182                           Distributed Model Predictive Control for Plant-Wide Systems

                                                             n
           Lemma 8.5 Suppose Assumptions 8.1–8.3 hold, x(k )∈ ℝ x , and conditions (8.18) and
                                                       0
           (8.20) are satisfied. For any k ≥ 0, if Problem 8.1 has a solution at every update time l,
                               f
           l = 1, 2, … ,k − 1, then x (k + N|k)∈Ω(  ∕2), ∀i ∈ P.
                               i
             Proof. In view of Lemmas 8.2 and 8.3, using the triangle inequality, we have
                             ‖ f       ‖    ‖ f                   ‖
                                                          i
                             ‖x (k + N|k)‖ ≤ ‖x (k + N|k) − ̂ x (k + N|k)‖
                                              i
                               i
                             ‖         ‖P i  ‖                    ‖P i
                                            + ̂ x (k + N|k) ‖
                                              ‖
                                              ‖ i
                                                        ‖P i
                                                   (1 −   )      
                                                 +                               (8.31)
                                          ≤ √        √     = √
                                            2 m     2 m      2 m
           for each j ∈ P , i ∈ P. This shows that the terminal state constraint is satisfied. This completes
                       i
           the proof.
                                                            n
           Theorem 8.1 Suppose Assumptions 8.1–8.3 hold, x(k )∈ ℝ x and constraints (8.10), (8.11),
                                                      0
                                                                  f
                                                                                f
           and (8.13) are satisfied at k . Then, for every i ∈ P, the control u (⋅|k) and state x (⋅|k),
                                  0
                                                                  i
                                                                                j,i
           respectively, defined by (8.17) and (8.4), are a feasible solution of Problem 8.1 at every
           update k.
             Proof. We will prove the theorem by induction.
                                                        p       f
             First, consider the case of k = 1. The state sequence x (⋅|1)= x (⋅|1) trivially satisfies the
                                                        j,i     j,i
           dynamic equation (8.4), the stability constraint (8.12), and the consistency constraints (8.10)
           and (8.11).
             Observe that
                                        p        f
                                ̂ x (1|1)= x (1|0)= x (1|1)= x (1), i ∈ P,
                                i       i        i       i
           and that
                                                 p
                                      f
                                     x (1 + s|1)= x (1 + s|0)
                                      i          i
                                             s = 1, 2, … , N − 1
                   f
             Thus, x (N|1)∈Ω (  ∕2). By the invariance of Ω(  ) under the terminal controller and the
                   i        i
           conditions in Lemma 8.1, it follows that the terminal state and control constraints are also
           satisfied. This completes the proof of the case of k = 1.
                          p      f
             Now suppose u (⋅|l)= u (⋅|l) is a feasible solution for l = 1, 2, … , k − 1. We will show that
                          i      i
            f
           u (⋅|k) is a feasible solution at update k.
            i
                                                                        f
             As before, the consistency constraint (8.10) is trivially satisfied, and x (⋅|k) is the cor-
                                                                        i
           responding state sequence that satisfies the dynamic equation. Since there is a solution for
           Problem 8.1 at updates l = 1, 2, … , k − 1, Lemmas 8.2–8.5 can be invoked. Lemma 8.4 guar-
           antees control constraint feasibility. Lemma 8.5 shows that the terminal state constraint is
           satisfied and the proof of Theorem 8.1 is completed.
   203   204   205   206   207   208   209   210   211   212   213