Skip to main content
Log in

Switched-mode systems: gradient-descent algorithms with Armijo step sizes

  • Published:
Discrete Event Dynamic Systems Aims and scope Submit manuscript

Abstract

This paper concerns optimal mode-scheduling in autonomous switched-mode hybrid dynamical systems, where the objective is to minimize a cost-performance functional defined on the state trajectory as a function of the schedule of modes. The controlled variable, namely the modes’ schedule, consists of the sequence of modes and the switchover times between them. We propose a gradient-descent algorithm that adjusts a given mode-schedule by changing multiple modes over time-sets of positive Lebesgue measures, thereby avoiding the inefficiencies inherent in existing techniques that change the modes one at a time. The algorithm is based on steepest descent with Armijo step sizes along Gâteaux differentials of the performance functional with respect to schedule-variations, which yields effective descent at each iteration. Since the space of mode-schedules is infinite dimensional and incomplete, the algorithm’s convergence is proved in the sense of Polak’s framework of optimality functions and minimizing sequences. Simulation results are presented, and possible extensions to problems with dwell-time lower-bound constraints are discussed.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7
Fig. 8
Fig. 9
Fig. 10

Similar content being viewed by others

Notes

  1. Note that the argument for proving Eq. 46 requires the assumption that \(v\in \mathcal {V}\), and may break down without it. Also, observe that it is possible that, for arbitrarily-large j, a mode present in v k(j) at a time s may not be present in v σ at any time near s; consequently a sharp inequality in Eq. 46 is possible, namely \(\lim _{j\rightarrow \infty }D_{\sigma _{k(j)}}<D_{\sigma }\).

References

  • Almér S, Mariéthoz S, Morari M (2012) Robust tracking control of pulse-width modulated systems through multi-frequency averaging. Proceedings ACC, Montreal, Canada, June 27–29

  • Almér S, Mariéthoz S, Morari M (2013) Sampled data model predictive control of a voltage source inverter for reduced harmonic distortion. IEEE Control Syst Technol. 21:1907-1915

  • Attia S A, Alamir M, Canudas de Wit C (2005) Sub optimal control of switched nonlinear systems under location and switching constraints. Proc. 16th IFAC World Congress, Prague, the Czech Republic, July 3–8

  • Axelsson H, Wardi Y, Egerstedt M, Verriest E (2008) A gradient descent approach to optimal mode scheduling in hybrid dynamical systems. J Optim Theory Appl 136:167–186

    Article  MATH  MathSciNet  Google Scholar 

  • Bengea S C, DeCarlo R A (2005) Optimal control of switching systems. Autom 41:11–27

    MATH  MathSciNet  Google Scholar 

  • Branicky M S, Borkar V S, Mitter S K (1998) A unified framework for hybrid control: model and optimal control theory. IEEE Trans Autom Control 43:31–45

    Article  MATH  MathSciNet  Google Scholar 

  • Brockett R (1995) Stabilization of motor networks. IEEE Conf Decision Control 1484–1488

  • Caines P, Shaikh M S (2005) Optimality zone algorithms for hybrid systems computation and control: exponential to linear complexity. Proceedings 13th Mediterranean conference on control and automation, Limassol, Cyprus, pp. 1292–1297, June 27–29

  • Caldwell T M, Murphey T D (2010) An adjoint method for second-order switching time optimization. Proceedings 49th CDC, Atlanta, Georgia, December 15–17

  • Caldwell T M, Murphey T D (2011) Switching mode generation and optimal estimation with application to skid-steering. Autom 47:50–64

    Article  MATH  MathSciNet  Google Scholar 

  • Caldwell T M, Murphey T D (3012) Projection-based switched system optimization: absolute continuity of the line search. Proceedings 51st CDC, Maui, Hawaii, December 10–13

  • Cassandras C G, Pepyne D L, Wardi Y (2001) Optimal control for a class of hybrid systems. IEEE Trans Autom Control 46 (3):398–415

    Article  MATH  MathSciNet  Google Scholar 

  • Ding X C, Wardi Y, Taylor D, Egerstedt M (2008) Optimization of switched-mode systems with switching costs. proceedings american control conference, Seattle, Washington, June 11–13

  • Egerstedt M (2000) Behavior based robotics using hybrid automata. Lecture notes in computer science: hybrid systems iii: computation and control. Springer Verlag, Pittsburgh, pp 103–116

    Google Scholar 

  • Egerstedt M, Wardi Y, Axelsson H (2006). IEEE Trans Autom Control AC–51 (1):110–115

    Article  MathSciNet  Google Scholar 

  • Flaßkamp K, Murphey T, Ober–Blöbaum S (2012) Switching time optimization in discretized hybrid dynamical systems. Proceedings 51st CDC, Maui, Hawaii, December 10–13

  • Gear C W (1971) Numerical initial value problems in ordinary differential equations. Prentice Hall Series in Automatic Computation, Englewood Cliffs, New Jersey

  • Gokbayrak K, Cassandras C G (2000) Hybrid controllers for hierarchically decomposed systems. Proceedings 2000 hybrid systems control Conference, pp. 117–129

  • Gonzalez H, Vasudevan R, Kamgarpour M, Sastry S S, Bajcsy R, Tomlin C (2010) A numerical method for the optimal control of switched systems. Proceedings 49th CDC, Atlanta, Georgia, pp. 7519-7526, December 15–17

  • Hristu-Varsakelis D. (2001) Feedback control systems as users of shared network: communication sequences that guarantee stability. IEEE conference on decision and control, pp. 3631–3636, Orlando, FL

  • Lincoln B, Rantzer A (2001) Optimizing linear systems switching. IEEE conference on decision and control, pp. 2063–2068, Orlando, FL

  • Meyer R T, Zefran M, Decarlo R A (2012) Comparison of multi-parametric programming, mixed-integer programming, gradient descent based, hybrid minimum principle, and the embedding approach on six published hybrid optimal control examples. Private Communication

  • Piccoli B (1998) Hybrid systems and optimal control. Proceedings IEEE conference on decision and control, Tampa, Florida, pp. 13–18

  • Polak E, Wardi Y (1984) A study of minimizing sequences. SIAM J Control Optim 22 (4):599–609

    Article  MATH  MathSciNet  Google Scholar 

  • Polak E (1997) Optimization algorithms and consistent approximations. Springer-Verlag, New York

    MATH  Google Scholar 

  • Rehbinder H, Sanfirdson M (2000) Scheduling of a limited communication channel for optimal control. ieee conference on decision and control, Sidney, Australia

  • Shaikh M S, Caines P (2002) On trajectory optimization for hybrid systems: theory and algorithms for fixed schedules. IEEE conference on decision and control, Las Vegas, NV

  • Shaikh M S, Caines P E (2003) On the optimal control of hybrid systems: optimization of trajectories, switching times and location schedules. In: Proceedings of the 6th international workshop on hybrid systems: computation and control, pp. 466-481, Prague, The Czech Republic

  • Shaikh M S, Caines P E (2005) Optimality zone algorithms for hybrid systems computation and control: from exponential to linear complexity. Proceedings IEEE conference on decision and control/european control conference, pp. 1403-1408, Seville, Spain

  • Shaikh M S, Caines P E (2007) On the hybrid optimal control problem: theory and algorithms. IEEE Trans Autom Control 52:1587–1603

    Article  MathSciNet  Google Scholar 

  • Sussmann H J (1999) A maximum principle for hybrid optimal control problems. Proceedings of the 38th IEEE conference on decision and control, pp. 425-430, Phoenix, AZ

  • Sussmann H J (2000) Set-valued differentials and the hybrid maximum principle. IEEE Conf Decis Control 1:558–563

    Google Scholar 

  • Vasudevan R, Gonzalez H, Bajcsy R, Sastry S S (2013) Consistent approximations for the optimal control of constrained switched systems - Part 1: a conceptual algorithm, and Part 2: an implementable algorithm. J Control Optim 51:4663–4483 (Part 1) and 4484–4503 (Part 2)

    Google Scholar 

  • Wang L Y, Beydoun A, Cook J, Sun J, Kolmanovsky I (1997) Optimal hybrid control with applications to automotive powertrain systems. Control using logic-based switching of LNCIS, vol 222, pp 190–200. Springer-Verlag

  • Wardi Y, Egerstedt M (2011) Algorithm for optimal mode scheduling in switched systems. Technical memorandum, Georgia Institute of Technology, http://www.ece.gatech.edu/~magnus/OptSwitchTechnReport.pdf,

  • Wardi Y, Egerstedt M (2012) Algorithm for optimal mode scheduling in switched systems. Proceedings ACC, Montreal, Canada, June 27–29

  • Xu X, Antsaklis P (2002) Optimal control of switched autonomous systems. IEEE conference on decision and control, Las Vegas, NV

  • Xu X, Antsaklis P J (2002) Optimal control of switched systems via nonlinear optimization based on direct differentiations of value functions. Int J Control 75:1406–1426

    Article  MATH  MathSciNet  Google Scholar 

  • Zhu F, Antsaklis P J (2011) Optimal control of switched hybrid systems: a brief survey. Technical report of the ISIS group at the University of Notre Dame, ISIS-2011-003

Download references

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Y. Wardi.

Additional information

Research supported in part by the National Science Foundation under Grant CNS-1239225

Appendix

Appendix

Proof of Lemma 2.

The proof follows from Proposition 5.6.8 and Proposition 5.6.10 in Polak (1997), which are stated in a more general context. For detailed arguments, consider the following setting. Let f 1:R nR n, f 2:R nR n, and L : R nR be C 2 functions, and fix T > 0 and x 0R n. Given γ ∈ [0, T], define the vector field

$$F(x, t;\gamma):=\left\{ \begin{array}{ll} f_{1}(x(t)), & \text{if}\ t\in[0, \gamma]\\ f_{2}(x(t)), & \text{if}\ t\in(\gamma, T], \end{array} \right.$$

and consider the differential equation \(\dot {x}(t))=F(x(t), t;\gamma )\) on the interval t ∈ [0, T], with the initial condition x(0) = x 0. Since the equation depends on γ, we denote its solution by x(t;γ). Define the performance function J(γ) by \(J(\gamma ):={{\int }_{0}^{T}}L(x(t;\gamma )dt\). Define the costate variable p(t;γ) by the equation \(\dot {p}(t;\gamma )=-\left (\frac {\partial F}{\partial x}(x, t;\gamma )\right )^{\top }p(t;\gamma )-\left (\frac {\partial L}{\partial t}(x;\gamma )\right )^{\top }\), dot denoting derivative with respect to t, with the boundary condition p(T;γ) = 0.

In the context of this paper, Lemma 2 amounts to the assertion that J(γ) is C 2. Reference Egerstedt et al. (2006) proved (in a more general context) that J(γ) is C 1 and the first derivative is given by

$$J^{\prime}(\gamma)=p(\gamma;\gamma)^{\top}\left(f_{1}x(\gamma;\gamma)-f_{2}(x(\gamma;\gamma))\right)$$
(59)

(see Proposition 2 2 and Proposition 3.1 there). Now the partial derivatives \(\frac {\partial x}{\partial t}(\gamma ;\gamma )\) and \(\frac {\partial x}{\partial \gamma }(\gamma ;\gamma )\) generally do not exist, but the total derivative \(\frac {dx}{d\gamma }(\gamma ;\gamma )\) exists and it is continuous. To see this note that for all t ∈ [0, γ] x(t;γ) satisfies the equation \(\dot {z}(t)=f_{1}(z(t))\), and hence \(\frac {dx}{d\gamma }(\gamma ;\gamma )=\dot {z}(\gamma )=f_{1}(x(\gamma ;\gamma )\), and the latter term is continuous by the assumption that f 1(x) is C 2. In a similar way, the total derivative \(\frac {dp}{d\gamma }(\gamma ;\gamma )\) exists and it is continuous. To see this, (Polak 1997) (Corollary 5.6.9) proves that for every t ∈ (γ, t] the derivative term \(\frac {\partial x}{\partial t}x(t;\gamma )\) is continuously differentiable in γ. Furthermore, by the costate equation the evolution of p(t;γ) backwards in time depends only on the vector field f 2 but not on f 1, and therefore \(\frac {dp}{d\gamma }(\gamma ;\gamma )=\left (-\left (\frac {\partial F}{\partial x}(x, \gamma ;\gamma )\right )^{\top }p(\gamma ;\gamma )-\left (\frac {\partial L}{\partial t}(x;\gamma )\right )^{\top }\right )^{\prime }\), “prime” denoting derivative with respect to γ; continuity follows by standard variational arguments on differentiability of differential equations (e.g., Corollary 5.6.9 in Polak (1997)) and the C 2 assumptions on f 1, f 2, and L. All of this implies that term in the the RHS of Eq. 59 is continuously differentiable thereby ascertaining that J(γ) is twice continuously differentiable. □

Proof of Lemma 3.

Consider a set S ⊂ [0, T) and schedules σ 1 and σ 2 as in the statement of the lemma. For i = 1, 2, let x i (⋅) and p i (⋅) denote the state trajectory and costate trajectory, respectively, associated with \(v_{\sigma _{i}}\). By Eqs. 1 and 5, and by Lemma 5.6.7 in Polak (1997) concerning Lipschitz continuity of solutions of differential equations, there exists K 1 > 0 such that, for all S ⊂ [0, T), σ 1 ∈ Σ, and σ 2 ∈ Σ as above,

$$||x_{1}-x_{2}||_{L^{\infty}}\ \leq\ K_{1}||v_{\sigma_{1}}-v_{\sigma_{2}}||_{L^{2}},$$
(60)

and

$$||p_{1}-p_{2}||_{L^{\infty}}\ \leq\ K_{1}||v_{\sigma_{1}}-v_{\sigma_{2}}||_{L^{2}}.$$
(61)

For every s ∈ [0, T)∖S, \(v_{\sigma _{1}}(s)=v_{\sigma _{2}}(s)\), and therefore, and by Eq. 6, for every wV,

$$\begin{array}{@{}rcl@{}} D_{\sigma_{1}, s, w}-D_{\sigma_{2}, s, w} \\ =\ p_{1}(s)^{T}\left(f(x_{1}(s), w)-f(x_{1}(s), v_{\sigma_{1}}(s))\right)- p_{2}(s)^{T}\left(f(x_{2}(s), w)-f(x_{2}(s), v_{\sigma_{1}}(s))\right). \end{array}$$
(62)

Consequently, and by Eqs. 60 and 61, there exists K 2 > 0 such that, for every S ⊂ [0, T), σ 1 ∈ Σ, and σ 2 ∈ Σ as above, and for every wV,

$$|D_{\sigma_{1}, s, w}-D_{\sigma_{2}, s, w}|\ \leq\ K_{2}||v_{\sigma_{1}}-v_{\sigma_{2}}||_{L^{2}}.$$
(63)

Since V is a finite set, and since \(v_{\sigma _{1}}(\tau )=v_{\sigma _{2}}(\tau )\)τ ∈ [0, T)∖S, there exists K 3 > 0 such that for all S ⊂ [0, T) and σ 1 ∈ Σ and σ 2 ∈ Σ as above, \(||v_{\sigma _{1}}-v_{\sigma _{2}}||_{L^{2}}\leq K_{3}\mu (S)\). This, together with Eq. 63, implies Eq. 22 with K := K 2 K 3. □

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Wardi, Y., Egerstedt, M. & Hale, M. Switched-mode systems: gradient-descent algorithms with Armijo step sizes. Discrete Event Dyn Syst 25, 571–599 (2015). https://doi.org/10.1007/s10626-014-0198-2

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s10626-014-0198-2

Keywords

Navigation