Chapter Summary, Key Terms, and Problems

Chapter summary¶

• Active engineering systems combine a plant, sensors, controller, actuators, and environment.

• Architecture, plant, and control are three distinct design domains—which components exist and how they connect, their continuous sizing, and how control action is generated—and the same premature-fixing hazard that motivates CCD at the plant-control level applies one level up, to architecture.

• Sequential design fixes the plant before the final controller is known and can eliminate better plant-controller combinations; the Millennium Bridge and combine-harvester header-height cases show this is a documented engineering failure mode, not a hypothetical one.

• CCD coordinates physical and control decisions through shared dynamics, objectives, information assumptions, and constraints, and is a specialization of the broader MDSDO/MDO landscape.

• CCD itself spans three complementary methodologies—control-inspired reasoning, mathematical co-optimization, and co-simulation—that this course develops primarily through co-optimization.

• Plant-control coupling exists when changing the plant changes the best controller and changing control capability changes the best plant; the five plant-objective cases explain precisely how an incomplete objective can hide this dependence.

• Iterated sequential design is a block-coordinate-descent algorithm; it can reach the CCD optimum only when both subproblems use a consistent, complete objective, and even then convergence slows as plant-control coupling strengthens.

• Passive and active elements can substitute for or complement one another, but differ in reliability, information, energy, and implementation requirements.

• CCD is especially valuable when dynamic performance is central, control authority is meaningful, constraints depend on trajectories, and architecture remains flexible.

• Open-loop optimal control gives a performance benchmark, not a deliverable; the gap to an implementable, limited-information controller can be an order of magnitude and is governed by how much information the controller is assumed to have (complete, instantaneous, or limited horizon).

• A lower objective value is meaningful only within a credible model, formulation, numerical solution, and implementation pathway.

Key terms¶

Problems¶

1. Quantifying plant--control coupling. An actively controlled oscillator satisfies $m\ddot x+c\dot x+kx=u+w$ and uses $u=-K_px-K_d\dot x$. For $J=\int_0^\infty(q_xx^2+q_v\dot x^2+r u^2)\,dt+\gamma_m m+gamma_k k$, derive a local cross-sensitivity measure that quantifies coupling between $(m,k)$ and $(K_p,K_d)$ at a stable design and explain how it predicts the potential value of CCD.

2. Sequential versus simultaneous design. For the static surrogate $J(p,c)=\tfrac12a(p-p_0)^2+bpc+\tfrac12d(c-c_0)^2$ with $a,d>0$ and $ad>b^2$, derive the exact simultaneous optimizer and the optimizer produced by one plant-then-control sequential pass, then obtain a closed-form expression for their objective gap.

3. Control-authority allocation. A positioning system has effective stiffness $k+K_p=q>0$ and cost $C(k,K_p)=a k^2+bK_p^2+\gamma kK_p$, with $k\ge0$ and $0\le K_p\le\bar K_p$. Derive the globally optimal passive--active allocation as a piecewise function of $(a,b,\gamma,q,\bar K_p)$.

4. A system-level objective from physical units. A battery-electric vehicle has longitudinal dynamics $m\dot v=F_t-\tfrac12\rho C_dA v^2-C_rm g-mg\sin\theta$ and battery power $P_b=F_tv/(\eta_d\eta_m)$. Construct one dimensionally consistent Bolza objective that trades trip time, electrical energy, battery mass, and terminal-speed error, and justify a normalization that makes its weights interpretable.

5. Architecture screening under common assumptions. Compare passive, semi-active, and fully active suspension architectures for $m_s\ddot z_s=-k_s(z_s-z_u)-c_s(\dot z_s-\dot z_u)+F_a$ under the same road input, packaging envelope, ride metric, and actuator-power model by formulating a single mixed-discrete CCD problem whose feasible sets make the comparison fair.

6. Failure of an incomplete plant objective. Let the true objective be $J(p,c)=(p-1)^2+(c-p)^2+\rho c^2$, while a plant team minimizes only $J_p(p)=(p-1)^2$ before the control team selects $c$. Derive the resulting sequential design and the simultaneous CCD design, then determine for which $\rho>0$ the relative performance loss exceeds ten percent.

7. Passive safety as a coupled requirement. The actuator in $m\ddot x+c\dot x+kx=u$ may fail at an unknown time, after which $u=0$. Formulate a CCD problem that minimizes nominal closed-loop performance while guaranteeing a prescribed exponential decay rate after failure, expressing the passive-safety condition as a constraint on the plant parameters.

8. Information as a design variable. A vehicle suspension controller receives a preview $w(t+\tau)$ of road displacement at sensing cost $C_s(\tau)=c_0+c_1\tau^2$, while the best achievable closed-loop cost is $J^*(p,\tau)$. Formulate a value-of-information test based on $\partial J^*/\partial\tau$ that determines whether the optimal design uses zero, finite, or maximum available preview.

9. Iterated sequential design. For a twice continuously differentiable strongly convex objective $J(p,c)$, represent alternating exact minimization over $p$ and $c$ as block coordinate descent and derive a local linear convergence factor in terms of the Hessian blocks $H_{pp}$, $H_{pc}$, and $H_{cc}$.

10. CCD study definition. For a two-link robot with $M(q,p)\ddot q+C(q,\dot q,p)\dot q+g(q,p)=B(p)u$, formulate a reproducible control co-design study with link dimensions as plant variables and a feedback policy as the control design, giving one complete objective, physically meaningful constraints, an information pattern, and a sequential baseline in one unified mathematical statement.

References and further reading¶

1. Garcia-Sanz, M. (2019). Control co-design: An engineering game changer. Advanced Control for Applications: Engineering and Industrial Systems, 1(1), Article e18. DOI: 10.1002/adc2.18

2. Allison, J. T., & Herber, D. R. (2014). Multidisciplinary design optimization of dynamic engineering systems. AIAA Journal, 52(4), 691–710. DOI: 10.2514/1.J052182

3. Allison, J. T., Guo, T., & Han, Z. (2014). Co-design of an active suspension using simultaneous dynamic optimization. Journal of Mechanical Design, 136(8), Article 081003.

4. Herber, D. R., & Allison, J. T. (2019). Nested and simultaneous solution strategies for general combined plant and control design problems. Journal of Mechanical Design, 141(1), Article 011402. DOI: 10.1115/1.4040705

5. Bayat, S., Peterson, C., Lee, Y. H., Iori, J., & Allison, J. T. (2026). Advancing wind turbines through control co-design: An integrative review. Applied Energy, 416, Article 127951. ScienceDirect article

6. Deshmukh, A. P., Herber, D. R., & Allison, J. T. (2015). Bridging the gap between open-loop and closed-loop control in co-design: A framework for complete optimal plant and control architecture design. In 2015 American Control Conference (ACC) (pp. 4916–4922).

7. Bayat, S., & Allison, J. T. (2026). Control co-design with varying available information applied to vehicle suspensions. ASME Journal of Dynamic Systems, Measurement, and Control, 148(1), Article 011013. DOI: 10.1115/1.4069918

8. Herber, D. R. (2017). Advances in Combined Architecture, Plant, and Control Design (Doctoral dissertation). University of Illinois at Urbana-Champaign.