Theory and Applications of Optimal Control Problems with Time Delays Helmut Maurer University of Münster Applied Mathematics: Institute of Analysis and Numerics Université Pierre et Marie Curie, Paris, March 10, 2017 Helmut Maurer Theory and Applications of Optimal Control Problems with Tim What can you expect from this talk ? Helmut Maurer Theory and Applications of Optimal Control Problems with Tim Challenges for Optimal Control Problems with Delays Theory and Numerics for non-delayed optimal control problems with control and state constraints are well developed: 1 2 3 4 Necessary and sufficient conditions, Stability and sensitivity analysis, Numerical methods: Boundary value methods, Discretization and NLP, Semismooth Newton methods, Real-time control techniques for perturbed extremals. CHALLENGE: Establish similar theoretical and numerical methods for delayed (retarded) optimal control problems. Helmut Maurer Theory and Applications of Optimal Control Problems with Tim Overview 1 2 3 4 5 6 7 Case Study: Combination Therapies for Cancer (with Ledzewicz, Schättler, Klamka, Swierniak) Optimal Control Problems with Time Delays in State and Control Variables Minimum Principle for State-Constrained Control Problems Numerical Treatment: Discretize and Optimize (with L. Göllmann) A Non-Convex Academic Example with a State Constraint Case Study:Two-stage Continuous Stirred Tank Reactor (CSTR) Case Study: Optimal Control of a Tuberculosis Model with Time Delays (with C. Silva, D.F. Torres) Helmut Maurer Theory and Applications of Optimal Control Problems with Tim Combination Therapies of Cancer Tumour Anti-Angiogenesis: J. Folkman (1972) et al. State and control variables: p : primary tumour volume [mm3 ] q : carrying capacity, or endothelial support [mm3 ] u : anti-angiogenic agent v : chemotoxic agent Helmut Maurer Theory and Applications of Optimal Control Problems with Tim Combination Therapies of Cancer: Literature U. Ledzewicz and H. Schaettler: Antiangiogenic therapy in cancer treatment as an optimal control problem, SIAM Journal on Control and Optimization, 46, (2007), 1052–1079. (Monotherapy) Hahnfeldt et al model with Gompertzian Growth: U. Ledzewicz, H. Maurer, and H. Schättler, Optimal and suboptimal protocols for a mathematical model for tumor antiangiogenesis in combination with chemotherapy, Mathematical Biosciences 22, pp. 13–26 (2009). Ergun et al model with Gompertzian Growth: U. Ledzewicz, H. Maurer, and H. Schaettler, On optimal delivery of combination therapy for tumors, Mathematical Biosciences and Engineering, 8, (2011), 307–323. Hahnfeldt et al model with Logistic Growth: J. Klamka, H. Maurer and A. Swierniak: Local Controllability and Optimal Control for a Model of Combined Anticancer therapy with Control Delays, Math. Biosc. Eng. 14(1), 195–216 (2017). Helmut Maurer Theory and Applications of Optimal Control Problems with Tim Optimal Control Problem p : tumor volume, u : anti-angiogenic control, y : total amount of u, q : carrying capacity, v : chemotoxic control, z : total amount of v . Dynamics of the Hahnfeldt et al model ṗ(t) = G (p(t), q(t)) − ϕ p(t) v (t), q̇(t) = b p(t) − q(t) (d p(t)2/3 + µ + γ u(t) + η v (t)) ẏ (t) = u(t), ż(t) = v (t). Initial conditions: p(0) = p0 , q(0) = q0 , y (0) = 0, z(0) = 0. Growth functions Gompertzian Growth : G (p, q) = −ξ p ln(p/q) Logistic Growth : G (p, q) = ξ p (1 − p/q) Helmut Maurer Theory and Applications of Optimal Control Problems with Tim Control problem and parameters Control problem Minimize final tumor volume p(T ) subject to the dynamic constraints, the control constraints 0 ≤ u(t) ≤ umax , 0 ≤ v (t) ≤ vmax , and the constraints on the total amount of drugs y (T ) ≤ ymax , z(T ) ≤ zmax . PARAMETERS (obtained from mice): ξ = 0.084, b = 5.85, d = 0.00873, γ = 0.15, ϕ = 0.2, η = 0.05, µ = 0.02. BOUNDS: umax = 75, Helmut Maurer ymax = 300, vmax = 2, zmax = 10. Theory and Applications of Optimal Control Problems with Tim Monotherapy : only anti-angiogenic control u Ledzewicz, Schättler (2007): Gompertzian Growth G (p, q) = −ξpln(p/q) , free terminal time T . Compute singular control in feedback form: q − (µ + d p 2/3 ) . u = using (p, q) = γ1 ξ ln pq + b qp + 23 ξ db p1/3 control u tumor p and vasculature q 80 70 60 50 40 30 20 10 0 p q 14000 11000 8000 5000 2000 0 1 2 3 4 5 time t (days) 6 0 1 2 3 4 5 time t (days) 6 Optimal control is bang-singular-bang. Sufficient conditions by synthesis analysis or switching time optimization. Helmut Maurer Theory and Applications of Optimal Control Problems with Tim Approximation of bang-singular-bang control u Gompertzian Growth and free terminal time. t1 = 0.07386, for 0 ≤ t < t1 umax u = 46.08 uc for t1 ≤ t ≤ t2 u(t) = , c t2 = 6.463 0 for t2 < t ≤ T T = 6.615 control u control u 80 70 60 50 40 30 20 10 0 80 70 60 50 40 30 20 10 0 0 1 2 3 4 5 time t (days) p(T ) = 8533 6 0 1 2 3 4 5 time t (days) 6 p(T ) = 8541 SSC hold for the approximative control w.r.t. z = (t1 , t2 , uc ). Helmut Maurer Theory and Applications of Optimal Control Problems with Tim Combination therapy: anti-angiogenic u, chemotherapy v Ledzewicz, M., Schättler (2009): Gompertzian Growth, T is free. Compute singular control in feedback form: u = using (p, q, v ) = γ1 ξ ln qp + b qp + 23 ξ d q b p 1/3 − (µ + d p 2/3 ) + ϕ−η γ v. Solution for ymax = 300 and zmax = 10 (total amount of drugs): control u control v 80 70 60 50 40 30 20 10 0 tumor p and vasculature q 15000 2 9000 1 6000 0.5 3000 0 0 1 2 3 time t (days) 4 5 p q 12000 1.5 0 0 1 2 3 4 time t (days) 5 0 1 2 3 4 5 time t (days) Surprising fact: chemotherapy v always starts later (pruning). Helmut Maurer Theory and Applications of Optimal Control Problems with Tim Combination Therapy with Logistic Growth Klamka, Maurer, Swierniak (2015): Logistic Growth G (p, q) = ξp(1 − p/q), T is free. Controls u and v are bang-bang: there are no singular arcs ! Chemotherapy control : v (t) ≡ 2 for terminal time T = 5. control u and switching function φu φu 100 80 60 40 20 0 -20 tumor p and vasculature q 16000 p q 12000 8000 4000 0 0 1 2 3 time t (days) 4 5 0 1 2 3 4 time t (days) 5 SSC hold for bang-bang controls ! Helmut Maurer Theory and Applications of Optimal Control Problems with Tim Dynamics with control delays p u v y : : : : tumor volume, anti-angiogenic control, chemotoxic control, total amount of u, q : carrying capacity, delay du = 10.6, delay dv = 1.84, z : total amount of v . Dynamics of the Hahnfeldt et al model ṗ(t) = G (p(t), q(t)) − ϕ p(t) v (t − dv ), q̇(t) = b p(t) − q(t) (d p(t)2/3 + µ + γ1 u(t) + γ2 u(t − du ) +η v (t)) ), ẏ (t) = u(t), ( 0 ≤ u(t) ≤ umax , y (T ) ≤ ymax ) ż(t) = v (t), ( 0 ≤ v (t) ≤ vmax , z(T ) ≤ zmax ) Initial conditions: p(0) = p0 , q(0) = q0 , y (0) = 0, z(0) = 0. Objective Minimize final tumor volume p(T ) Helmut Maurer Theory and Applications of Optimal Control Problems with Tim Combination therapy with control delays Klamka, Maurer, Swierniak (2015) : Logistic Growth, T = 16 fixed. Solution for umax = 40, ymax = 300 and vmax = 2, zmax = 10 , Delays du = 10.6, dv = 1.84 control u and switching function φu φu 50 40 30 20 10 0 -10 -20 0 2 4 6 8 10 12 14 16 time t (days) control v and switching function φv 3 2.5 2 1.5 1 0.5 0 -0.5 φv tumor p and vasculature q 20000 p q 15000 10000 5000 0 0 2 4 6 8 10 12 14 16 0 2 time t (days) 4 6 8 10 12 14 16 time t (days) Necessary conditions are satisfied (extremal solution), but we cannot check sufficient conditions. Helmut Maurer Theory and Applications of Optimal Control Problems with Tim Delayed Optimal Control Problem with State Constraints State x(t) ∈ Rn , Control u(t) ∈ Rm , Delays dx , du ≥ 0. Dynamics and Boundary Conditions ẋ(t) = f (x(t), x(t − dx ), u(t), u(t − du )), a.e. t ∈ [0, tf ] , x(t) = x0 (t), t ∈ [−dx , 0] , u(t) = u0 (t), t ∈ [−du , 0), ψ(x(tf )) = 0q (0 ≤ q ≤ n). Control and State Constraints umin ≤ u(t) ≤ umax , S(x(t)) ≤ 0, ∀ t ∈ [0, tf ] (S : Rn → Rk ) . Minimize Z tf f0 (x(t), x(t − dx ), u(t), u(t − du )) dt J(u, x) = g (x(tf )) + 0 Helmut Maurer Theory and Applications of Optimal Control Problems with Tim Literature on optimal control with time-delays Time delays in state variables and pure control constraints: Kharatishvili (1961), Oguztöreli (1966), Banks (1968), Halanay (1968), Soliman, Ray (1970, chemical engineering), Warga (1968,1972): optimization in Banach spaces), Guinn (1976) : transform delayed problems to standard problems), Colonius, Hinrichsen (1978), Clarke, Wolenski (1991), Dadebo, Luus (1992), Mordukhovich, Wang (2003–). Time delays in state variables and pure state constraints: Angell, Kirsch (1990). State and control delays and mixed control–state constraints: Göllmann, Maurer (OCAM 2009, JIMO 2014), Time delays in state and control variables and state constraints: Vinter (2016) : Maximum Principle for a general problem Helmut Maurer Theory and Applications of Optimal Control Problems with Tim Method of steps : Warga (1968) and Guinn (1976) Transform an optimal control problem with delays to a standard non–delayed optimal control problem: requires commensurability. Then apply the necessary conditions for non-delayed problems: Jacobson, Lele, Speyer (1975): KKT conditions in Banach spaces. Maurer (1979) : Regularity of multipliers for state constraints. Hartl, Sethi, Thomsen (SIAM Review 1995): Survey on Maximum Principles. Vinter (2000): (Nonsmooth) Optimal Control. Applications to mixed control-state constraints: single delays: Göllmann, Kern, Maurer (OCAM 2009), multiple delays: Göllmann, Maurer (JIMO 2014) Helmut Maurer Theory and Applications of Optimal Control Problems with Tim Hamiltonian Hamiltonian (Pontryagin) Function H(x, y , λ, u, v ) :=λ0 f0 (t, x, y , u, v ) + λf (t, x, y , u, v ) y variable with y (t) = x(t − dx ) v variable with v (t) = u(t − du ) λ ∈ Rn , λ0 ∈ R adjoint (costate) variable Let (u, x) ∈ L∞ ([0, tf ], Rm ) × W 1,∞ ([0, tf ], Rn ) be a locally optimal pair of functions. Then there exist an adjoint function λ ∈ BV([0, tf ], Rn ) and λ0 ≥ 0, a multiplier ρ ∈ Rq (associated with terminal conditions), a multiplier function (measure) µ ∈ BV([0, tf ], Rk ), such that the following conditions are satisfied for a.e. t ∈ [0, tf ] : Helmut Maurer Theory and Applications of Optimal Control Problems with Tim Minimum Principle (i) Advanced adjoint equation and transversality condition: λ(t) = Rtf t ( Hx (s) + χ[0,tf −dx ] (s) Hy (s + dx ) ) ds + + (λ0 g + ρψ)x (x(tf )) Rtf Sx (x(s)) dµ(s) t ( if S(x(tf )) < 0 ), where Hx (t) and Hy (t + dx ) denote evaluations along the optimal trajectory and χ[0,tf −dx ] is the characteristic function. (ii) Minimum Condition for U = [ umin , umax ] : H(t) + χ[0,tf −du ] (t) H(t + du ) = min w ∈U [ H(x(t), y (t), λ(t), w , v (t)) + χ[0,tf −du ] (t) H(x(t + du ), y (t), λ(t + du ), u(t + du ), w ) ]. (iii) Multiplier condition and complementarity condition: Ztf dµ(t) ≥ 0, S(x(t)) dµ(t) = 0. 0 Helmut Maurer Theory and Applications of Optimal Control Problems with Tim Regularity conditions for dµ(t) = η(t)dt if du = 0 Boundary arc : S(x(t)) = 0 for t1 ≤ t ≤ t2 . Assumption : u(t) ∈ int(U) for t1 < t < t2 . Under certain regularity conditions we have dµ(t) = η(t) dt with a continuous multiplier η(t) for all t1 < t < t2 . Adjoint equation and jump conditions λ̇(t) = −Hx (t) − χ[0,tf −dx ] (t) Hy (t + dx ) − η(t)Sx (x(t)) λ(tk +) = λ(tk −) − νk Sx (x(tk )) , νk ≥ 0, at each contact or junction time tk , νk = µ(tk +) − µ(tk −). Minimum condition: no control constraints and delays Hu (t) = 0 . This condition allows to compute the multiplier η = η(x, λ). Helmut Maurer Theory and Applications of Optimal Control Problems with Tim Discretization and NLP Consider MAYER problem with cost functional J(u, x) = g (x(tf )) . Commensurability Assumption: There exists a stepsize h > 0 and integers k, l, N ∈ N with dx = k · h, Grid points ti := i · h du = l · h, tf = N · h . (i = 0, 1, . . . , N), tN = tf . Approximation at grid points: u(ti ) ≈ ui ∈ Rm , x(ti ) ≈ xi ∈ Rn (i = 0, . . . , N). For simplicity EULER method: xi+1 = xi + h · f (ti , xi , xi−k , ui , ui−l ), x−i := x0 (−ih) Helmut Maurer (i = 0, .., k), i = 0, 1, ..., N − 1, u−i := u0 (−ih) (i = 1, .., l). Theory and Applications of Optimal Control Problems with Tim Large-scale NLP Problem Include mixed control-state constraint C (x(t), u(t)) ≤ 0. Minimize J(u, x) = g (xN ) subject to xi+1 − xi − h · f (ti , xi , xi−k , ui , ui−l ) = 0, i = 0, .., N − 1, ψ(xN ) = 0, x−i := x0 (−ih) (i = 0, .., k), C (xi , ui ) ≤ 0, i = 0, .., N, S(xi ) ≤ 0, i = 0, .., N, u−i := u0 (−ih) (i = 1, .., l). Optimization Variable: z := (u0 , x1 , u1 , x2 , ..., uN−1 , xN ) ∈ RN(m+n) Helmut Maurer Theory and Applications of Optimal Control Problems with Tim Multipliers and NLP-Solvers Approximations: Adjoint variable : λ(ti ) ≈ λi Multiplier , multiplier for ODE, : η(ti ) ≈ ηi /h , multiplier for S(xi ) ≤ 0. AMPL : Programming language (Fourer, Gay, Kernighan) IPOPT: Interior point method (A. Wächter et al.) LOQO: Interior point method (B. Vanderbei et al. WORHP : SQP–method (C. Büskens, M. Gerdts) Other NLP solvers embedded in AMPL : cf. NEOS server. Special feature: solvers provide LAGRANGE-multipliers. BOCOP : F. Bonnans, P. Martinon. Helmut Maurer Theory and Applications of Optimal Control Problems with Tim Academic Example: state delays state x(t) ∈ R, control u(t) ∈ R, delay d ≥ 0 Dynamics and Boundary Conditions ẋ(t) = x(t − d)2 − u(t), x(t) = x0 (t) = 1, t ∈ [0, 2], t ∈ [−d, 0], x(2) = 1 Control and State Constraints x(t) ≥ α, i.e., S(x(t)) = −x(t) + α ≤ 0, t ∈ [0, 2] Minimize Z J(u, x) = 2 (x(t)2 + u(t)2 ) dt 0 Helmut Maurer Theory and Applications of Optimal Control Problems with Tim Optimal solutions without state constraints Z 2 min (x(t)2 +u(t)2 ) dt s.t. ẋ(t) = x(t − d)2 −u(t), x0 (t) ≡ 1, x(2) = 1 0 optimal state and control for delays d = 0.0, d = 0.1, d = 0.2, state x 1 0.95 0.9 0.85 0.8 0.75 0.7 0.65 0.6 0.55 0.5 d = 0.5, control u 2.5 d=0 d=0.1 d=0.2 d=0.5 2 1.5 d=0 d=0.1 d=0.2 d=0.5 1 0.5 0 -0.5 0 Helmut Maurer 0.5 1 1.5 2 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 Theory and Applications of Optimal Control Problems with Tim Optimal solutions with state constraint x(t) ≥ α = 0.7 Optimal state and control for delays d = 0.0, d = 0.1, d = 0.2, state x 1 control u 2.5 d=0 d=0.1 d=0.2 d=0.5 0.95 0.9 d = 0.5 d=0 d=0.1 d=0.2 d=0.5 2 1.5 0.85 1 0.8 0.5 0.75 0.7 0 0.65 -0.5 0 0.5 1 1.5 2 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 Optimal controls u(t) are continuous ! Helmut Maurer Theory and Applications of Optimal Control Problems with Tim Minimum Principle Augmented Hamiltonian: y (t) = x(t − d) H(x, y , λ, η, u) = u 2 + x 2 + λ(y 2 − u) + η(−x + α) Adjoint equation λ̇(t) = −Hx (t) − χ[0,2−d] Hy (t + d) −2x(t) − 2λ(t + d)x(t) + η(t) , 0 ≤ t ≤ 2 − d = −2x(t) + η(t) , 2−d ≤t ≤2 Minimum condition Hu (t) = 0 Helmut Maurer ⇒ u(t) = λ(t)/2 Theory and Applications of Optimal Control Problems with Tim Boundary arc x(t) = α = 0.7 for t1 ≤ t ≤ t2 x(t) ≡ α ⇒ x(t − d)2 = u(t) = λ(t)/2 ⇒ ẋ(t) = 0 Computation of multiplier η(t) by differentiation η(t) = 2(2x(t −d)(x(t −2d)2 −λ(t −d)/2)+x(t)+λ(t +d)x(t)) delays d = 0.0, d = 0.1, d = 0.2, 1 0.9 d = 1.0 3 d=0 d=0.1 d=0.2 d=0.5 0.95 d = 0.5, multiplier η state x 2.5 2 0.85 1.5 0.8 1 0.75 0.5 0.7 0 0.65 0 Helmut Maurer 0.5 1 1.5 2 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 Theory and Applications of Optimal Control Problems with Tim Two-Stage Continuous Stirred Tank Reactor (CSTR) Time delays are caused by transport between the two tanks. Helmut Maurer Theory and Applications of Optimal Control Problems with Tim Two Stage CSTR Dadebo S., Luus R. Optimal control of time-delay systems by dynamic programming, Optimal Control Applications and Methods 13, pp. 29–41 (1992). A chemical reaction A ⇒ B is processed in two tanks. State and control variables: Tank 1 : Tank 2 : x1 (t) : (scaled) concentration x2 (t) : (scaled) temperature u1 (t) : temperature control x3 (t) : (scaled) concentration x4 (t) : (scaled) temperature u2 (t) : temperature control State variables denote deviations from equilibrium. Helmut Maurer Theory and Applications of Optimal Control Problems with Tim Dynamics of the Two-Stage CSTR Reaction term in Tank 1 : R1 (x1 , x2 ) = (x1 + 0.5) exp Reaction term in Tank 2 : R2 (x3 , x4 ) = (x3 + 0.25) exp 25x2 1+x2 25x4 1+x4 Dynamics: ẋ1 (t) = −0.5 − x1 (t) − R1 (t), ẋ2 (t) = −(x2 (t) + 0.25) − u1 (t)(x2 (t) + 0.25) + R1 (t), ẋ3 (t) = x1 (t − d) − x3 (t) − R2 (t) + 0.25, ẋ4 (t) = x2 (t − d) − 2x4 (t) − u2 (t)(x4 (t) + 0.25) + R2 (t) − 0.25. Initial conditions: x1 (t) = 0.15, x2 (t) = −0.03, −d ≤ t ≤ 0, x3 (0) = 0.1, x4 (0) = 0. Delays d = 0.1, d = 0.2, d = 0.4 in the state variables x1 , x2 . Helmut Maurer Theory and Applications of Optimal Control Problems with Tim Optimal control problem for the Two-Stage CSTR Minimize Rtf ( x12 + x22 + x32 + x42 + 0.1u12 + 0.1u22 ) dt (tf = 2) . 0 Hamiltonian with yk (t) = xk (t − d), k = 1, 2 : H(x, y , λ, u) = f0 (x, u) + λ1 ẋ1 +λ2 (−(x2 + 0.25) − u1 (x2 + 0.25) + R1 (x1 , x2 ) ) +λ3 (y1 − x3 − R2 (x3 , x4 ) + 0.25) +λ4 (y2 − 2x4 − u2 (x4 + 0.25) + R2 (x3 , x4 ) + 0.25) Advanced adjoint λ̇1 (t) λ̇2 (t) λ̇k (t) equations: = −Hx1 (t) − χ [ 0,tf −d ] (t) λ3 (t + d), = −Hx2 (t) − χ [ 0,tf −d ] (t) λ4 (t + d), = −Hxk (t) (k = 3, 4). The minimum condition yields Hu = 0 and thus u1 = 5λ2 (x2 + 0.25), Helmut Maurer u2 = 5λ4 (x4 + 0.25). Theory and Applications of Optimal Control Problems with Tim Two-Stage CSTR with free x(tf ) : x1 , x2 , x3 , x4 concentration x1 0.16 0.14 0.12 0.1 0.08 0.06 0.04 0.02 0 -0.02 temperature x2 d=0.1 d=0.2 d=0.4 d=0.1 d=0.2 d=0.4 0.02 0.01 0 -0.01 -0.02 -0.03 0 0.5 1 1.5 2 0 0.5 concentration x3 0.1 0.06 0.02 0 -0.02 -0.04 0.5 1 1.5 Delays Helmut Maurer 2 d=0.1 d=0.2 d=0.4 0.045 0.04 0.035 0.03 0.025 0.02 0.015 0.01 0.005 0 0.04 0 1.5 temperature x4 d=0.1 d=0.2 d=0.4 0.08 1 2 0 0.5 1 1.5 2 d = 0.1, d = 0.2, d = 0.4. Theory and Applications of Optimal Control Problems with Tim Two-Stage CSTR with free x(tf ) : u1 , u2 , λ2 , λ4 control u1 control u2 0.3 d=0.1 d=0.2 d=0.4 0.2 0.1 d=0.1 d=0.2 d=0.4 0.25 0.2 0 0.15 -0.1 -0.2 0.1 -0.3 0.05 -0.4 0 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 adjoint variable λ2 0.3 adjoint variable λ4 0.2 0.18 0.16 0.14 0.12 0.1 0.08 0.06 0.04 0.02 0 d=0.1 d=0.2 d=0.4 0.2 0.1 0 -0.1 -0.2 -0.3 -0.4 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 Delays Helmut Maurer d=0.1 d=0.2 d=0.4 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 d = 0.1, d = 0.2, d = 0.4. Theory and Applications of Optimal Control Problems with Tim Two-Stage CSTR with x(tf ) = 0 : x1 , x2 , x3 , x4 concentration x1 0.16 0.14 0.12 0.1 0.08 0.06 0.04 0.02 0 temperature x2 0.02 d=0.1 d=0.2 d=0.4 d=0.1 d=0.2 d=0.4 0.01 0 -0.01 -0.02 -0.03 0 0.5 1 1.5 2 0 0.5 concentration x3 0.1 1.5 2 temperature x4 0.035 d=0.1 d=0.2 d=0.4 0.08 1 d=0.1 d=0.2 d=0.4 0.03 0.025 0.06 0.02 0.04 0.015 0.02 0.01 0 0.005 -0.02 0 0 0.5 1 1.5 Delays Helmut Maurer 2 0 0.5 1 1.5 2 d = 0.1, d = 0.2, d = 0.4. Theory and Applications of Optimal Control Problems with Tim Two-Stage CSTR with x(tf ) = 0 : u1 , u2 , λ3 , λ4 control u1 control u2 0.35 d=0.1 d=0.2 d=0.4 0.2 0.1 d=0.1 d=0.2 d=0.4 0.3 0.25 0 0.2 -0.1 0.15 -0.2 0.1 -0.3 0.05 -0.4 0 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 adjoint variable λ3 0.15 adjoint variable λ4 0.25 d=0.1 d=0.2 d=0.4 0.1 d=0.1 d=0.2 d=0.4 0.2 0.05 0.15 0 0.1 -0.05 0.05 -0.1 0 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 Delays Helmut Maurer 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 d = 0.1, d = 0.2, d = 0.4. Theory and Applications of Optimal Control Problems with Tim Two-Stage CSTR with x(tf ) = 0 and x4 (t) ≤ 0.01 temperature x4 0.016 0.014 0.012 0.01 0.008 0.006 0.004 0.002 0 control u1 0.3 0.2 0.1 0 -0.1 -0.2 -0.3 -0.4 -0.5 d=0.1 d=0.2 d=0.4 0 0.5 1 1.5 2 d=0.1 d=0.2 d=0.4 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 multiplier η for x4 <= 0.01 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 control u2 0.35 d=0.1 d=0.2 d=0.4 d=0.1 d=0.2 d=0.4 0.3 0.25 0.2 0.15 0.1 0.05 0 0.5 1 1.5 Delays Helmut Maurer 2 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 d = 0.1, d = 0.2, d = 0.4. Theory and Applications of Optimal Control Problems with Tim Two-Stage CSTR: time-optimal with x(tf ) = 0 d = 0 : tf = 1.8725496 control u1 and (scaled) switching function φ1 control u1 0.4 0.3 0.2 0.1 0 -0.1 -0.2 -0.3 -0.4 -0.5 -0.6 0.2 0.15 0.1 0.05 0 -0.05 -0.1 -0.15 -0.2 0 0.5 1 1.5 2 u1 φ1 0 0.5 1 0.4 0.3 0.2 0.1 0 -0.1 -0.2 -0.3 -0.4 0.2 0.15 0.1 0.05 0 -0.05 -0.1 -0.15 -0.2 0 0.5 1 1.5 2 control u1 and (scaled) switching function φ1 control u1 1.5 2 u1 φ1 0 0.5 1 1.5 2 d = 0.4 : tf = 2.0549 Boccia, Falugi, Maurer, Vinter : CDC 2013 . Vinter (2016) : Maximum Principle in general form. Helmut Maurer Theory and Applications of Optimal Control Problems with Tim Compartment Model for Tuberculosis C.J. Silva and D.F.M. Torres, Optimal control strategies for Tuberculosis treatment: a case study in Angola, Numerical Algebra, Control and Optimization, 2 (2012), 601–617. C.J. Silva, H. Maurer, and D.F.M. Torres, Optimal control of a Tuberculosis model with state and control delays, Mathematical Biosciences and Engineering 14(1), pp. 321–337 (2017). 5 state variables and delay in I (t): S L1 I L2 R N : : : : : : Susceptible individuals early latent individuals, recently infected (less than two years) infectious individuals, who have active TB persistent latent individuals recovered individuals total population N = S + L1 + I + L2 + R, assumed constant dI : delay in I , represents delay in diagnosis Helmut Maurer Theory and Applications of Optimal Control Problems with Tim Dynamical TB model with state and control delays Control variables and delays: u1 : effort on early detection and treatment of recently infected individuals L1 , du1 : delay on the diagnosis of latent TB and commencement of latent TB treatment, u2 : chemotherapy or post-exposure vaccine to persistent latent individuals (L2 ), du2 : delay in the prophylactic treatment of persistent latent L2 . Dynamical model : R = N − S − L1 − I − L2 Ṡ(t) = µN − L̇1 (t) = β N I (t)S(t) − µS(t), β N I (t)(S(t) + σL2 (t) + σR R(t)) −(δ + τ1 + 1 u1 (t − du1 ) + µ)L1 (t), İ (t) = φδL1 (t) + ωL2 (t) + ωR R(t) − τ0 I (t − dI ) − µI (t), L̇2 (t) = (1 − φ)δL1 (t) − σ Nβ I (t)L2 (t) −(ω + 2 u2 (t − du2 ) + τ2 + µ)L2 (t). Helmut Maurer Theory and Applications of Optimal Control Problems with Tim Optimal control problem with state and control delays Initial functions in view of delays (N = 30000) : S(0) = 76 N/120, L1 (0) = 36 N/120, L2 (0) = 2 N/120, I (t) = 5 N/120 ∀ − dI ≤ t ≤ 0, R(0) = N/120, uk (t) = 0 ∀ − duk ≤ t < 0, k = 1, 2. Control constraints 0 ≤ u1 (t) ≤ u1max = 1, 0 ≤ u2 (t) ≤ u2max = 1 for 0 ≤ t ≤ T . Control problem: x = (S, L1 , I , L2 ) ∈ R4 , u = (u1 , u2 ) ∈ R2 Minimize L1 objective or L2 objective RT J1 (x, u) = 0 ( I (t) + L2 (t) + W1 u1 (t) + W2 u2 (t) ) dt , RT J2 (x, u) = 0 ( I (t) + L2 (t) + W1 u1 (t)2 + W2 u2 (t)2 ) dt (weights W1 , W2 > 0) subject to the dynamic and control constraints. Helmut Maurer Theory and Applications of Optimal Control Problems with Tim Parameters in the TB Model Symbol β µ δ φ ω ωR σ σR τ0 τ1 τ2 N T 1 2 Description Transmission coefficient Death and birth rate Rate at which individuals leave L1 Proportion of individuals going to I Reactivation rate for persistent latent infections Reactivation rate for treated individuals Factor reducing the risk of infection as a result of aquired immunity to a previous infection for L2 Rate of exogenous reinfection of treated patients Rate of recovery under treatment of active TB Rate of recovery under treatment of L1 Rate of recovery under treatment of L2 Total population Total simulation duration Efficacy of treatment of early latent L1 Efficacy of treatment of persistent latent L2 Value 100 1/70 yr −1 12 yr −1 0.05 0.0002 yr −1 0.00002 yr −1 0.25 0.25 2 yr −1 2 yr −1 1 yr −1 30, 000 5 yr 0.5 0.5 Tabelle: Parameter values. Helmut Maurer Theory and Applications of Optimal Control Problems with Tim Optimal controls of the non-delayed control problems Adjoint variable λ = (λS , λL1 , λI , λL2 ) Switching functions φk (t) = Huk [t] = −Wk − k λLk (t) Lk (t) (k = 1, 2). Control law for Maximum Principle: 1 , if φk (t) > 0 uk (t) = , k = 1, 2. 0 , if φk (t) < 0 control u1 and switching function φ1 2 u1 φ1 1.5 control u2 and switching function φ2 3 2.5 u2 φ2 2 1 1.5 0.5 0.5 1 0 0 -0.5 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 Weights W1 = W2 = 50 : Optimal controls are bang-bang. SSC are satisfied, since SSC hold for the switching time optimization problem and the strict bang-bang property holds. Helmut Maurer Theory and Applications of Optimal Control Problems with Tim State Variables in the Non-Delayed Control Problem Optimal state variables for weights W1 = W2 = 50 : susceptible S and recovered R infectious I 1600 1400 1200 1000 800 600 400 200 0 S R 35000 30000 25000 20000 15000 10000 5000 0 0 1 2 3 4 5 0 1 early latent L1 9000 8000 7000 6000 5000 4000 3000 2000 1000 0 3 4 5 4 5 14000 12000 10000 8000 6000 4000 2000 0 0 Helmut Maurer 2 persistent latent L2 1 2 3 4 5 0 1 2 3 Theory and Applications of Optimal Control Problems with Tim Optimal controls : comparison of L1 and L2 objectives Objectives: J1 (x, u) = RT J2 (x, u) = RT 0 0 ( I (t) + L2 (t) + W1 u1 (t) + W2 u2 (t) ) dt, ( I (t) + L2 (t) + W1 u1 (t)2 + W2 u2 (t)2 ) dt L2 functional : optimal controls are continuous. control u1 for J1 and J2 objective 1.4 1.2 1 0.8 0.6 0.4 0.2 0 J1 J2 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 control u2 for J1 and J2 objective 1.4 1.2 1 0.8 0.6 0.4 0.2 0 J1 J2 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 Weights W1 = W2 = 50 Helmut Maurer Theory and Applications of Optimal Control Problems with Tim Optimal controls for delays dI = 0.1, du1 = du2 = 0.2 Switching functions for k = 1, 2 ) ( −Wk − k λL1 (t + duk )L1 (t + duk ) for 0 ≤ t ≤ T − duk φk (t) = −Wk for T − duk ≤ t ≤ T Optimal controls u1 and u2 are bang-bang with one switch. The strict bang-bang property holds. control u1 and (scaled) switching function φ 1.2 u1 1 φ1 0.8 0.6 0.4 0.2 0 -0.2 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 susceptibles S and recovered R 0 1 control u2 and (scaled) switching function φ 3 2.5 2 1.5 1 0.5 0 -0.5 u2 φ2 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 Helmut Maurer 2 3 infectious individuals I 1600 1400 1200 1000 800 600 400 200 0 S R 35000 30000 25000 20000 15000 10000 5000 0 4 5 0 1 early latent L1 2 3 4 5 4 5 early latent L2 9000 8000 7000 6000 5000 4000 3000 2000 1000 0 14000 12000 10000 8000 6000 4000 2000 0 0 1 2 3 4 5 0 1 2 3 Theory and Applications of Optimal Control Problems with Tim Sensitivity of optimal control w.r.t. parameter β control u1 1.4 1.2 1 0.8 0.6 0.4 0.2 0 β=50 β=150 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 control u2 1.4 1.2 1 0.8 0.6 0.4 0.2 0 β=150 β=150 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 Comparison of extremal controls for β = 50 and β = 150 : L1 objective, W1 = W2 = 150, delays dI = 0.1, du1 = du2 = 0.2. Helmut Maurer Theory and Applications of Optimal Control Problems with Tim Trahison des Images ? Ceci n’était pas une conférence. Merci pour votre attention ! Helmut Maurer Theory and Applications of Optimal Control Problems with Tim
© Copyright 2026 Paperzz