Differential Games, Distributed Systems, and Impulse Control
|
|
- Shonda Gallagher
- 6 years ago
- Views:
Transcription
1 Chapter 12 Differential Games, Distributed Systems, and Impulse Control In previous chapters, we were mainly concerned with the optimal control problems formulated in Chapters 3 and 4 and their applications to various functional areas of management and to some problems in economics. These problems were described by a single objective function (or a single decision maker) and a set of ordinary differential equations, called the state equations, defined in a deterministic framework. In this chapter, we deal with generalizations of the (ordinary) deterministic optimal control problems that can be made in one or more of the following directions. Generalizations involving stochastic optimal control problems are discussed in the next chapter. There may be more than one decision maker, each having one s own objective function that each is trying to maximize, subject to a set of differential equations. This extension of the optimal control theory is referred to as the theory of differential games. Section 12.1 contains a brief introduction to differential games along with an application. Another extension replaces the system of ordinary differential equations by a set of partial differential equations. These come under the classification of distributed parameter systems and are treated in Section Finally in Section 12.3, we treat the theory of impulse control. These controls are developed to deal with systems which, in addition to conventional controls, allow a controller to make discrete changes in the state 37
2 12.1. Differential Games 39 where U and V are convex sets in E 1. Consider further the objective function T J(u, v) = S[x(T)] + F(x, u, v, t)dt, (12.2) which player 1 wants to maximize and player 2 wants to minimize. Since the gain of player 1 represents a loss to player 2, such games are appropriately termed zero-sum games. Clearly, we are looking for admissible control trajectories u and v such that J(u, v) J(u, v ) J(u, v ). (12.3) The solution (u, v ) is known as the minimax solution. Here u and v stand for u (t), t [, T], and v (t), t [, T], respectively. The necessary conditions for u and v to satisfy (12.3) are given by an extension of the maximum principle. To obtain these conditions, we form the Hamiltonian H = F + λf (12.4) with the adjoint variable λ satisfying the equation λ = H x, λ(t) = S x [x(t)]. (12.5) The necessary condition for trajectories u and v to be a minimax solution is that for t [, T], H(x (t), u (t), v (t), λ(t), t) = min v V max u U H(x (t), u, v,λ(t), t), (12.6) which can also be stated, with suppression of (t), as H(x, u, v,λ, t) H(x, u, v, λ, t) H(x, u, v, λ, t) (12.7) for u U and v V. Note that (u, v ) is a saddle point of the Hamiltonian function H. Note that if u and v are unconstrained, i.e., when, U = V = E 1, condition (12.6) reduces to the first-order necessary conditions and the second-order conditions are H u = and H v =, (12.8) H uu and H vv. (12.9) We now turn to the treatment of nonzero-sum differential games.
3 Differential Games, Distributed Systems, and Impulse Control Nonzero-Sum Differential Games In this section, let us assume that we have N players where N 2. Let u i U i, i = 1, 2,..., N, represent the control variable for the ith player, where U i is the set of controls from which the ith player can choose. Let the state equation be defined as Let J i, defined by J i = S i [x(t)] + ẋ = f(x, u 1, u 2,..., u N, t). (12.1) T F i (x, u 1, u 2,..., u N, t)dt, (12.11) denote the objective function which the ith player wants to maximize. In this case, a Nash solution is defined by a set of N admissible trajectories which have the property that {u 1, u 2,..., u N }, (12.12) J i (u 1, u 2,..., u N ) = max u i U Ji (u 1, u 2,..., u (i 1), u i,..., u (i+1),..., u N ) i (12.13) for i = 1, 2,..., N. To obtain the necessary conditions for a Nash solution for nonzerosum differential games, we must make a distinction between open-loop and closed-loop controls. Open-Loop Nash Solution The open-loop Nash solution is defined when (12.12) is given as functions of time satisfying (12.13). To obtain the maximum principle type conditions for such solutions to be a Nash solution, let us define the Hamiltonian functions H i = F i + λ i f (12.14) for i = 1, 2,..., N, with λ i satisfying λ i = H i x, λ i (T) = S i x[x(t)]. (12.15) The Nash control u i for the ith player is obtained by maximizing the ith Hamiltonian H i with respect to u i, i.e., u i must satisfy H i (x, u 1,..., u (i 1), u i, u (i+1),..., u N, λ, t) H i (x, u 1,..., u (i 1), u i, u (i+1),..., u N, λ, t), t [, T], (12.16)
4 12.1. Differential Games 311 for all u i U i, i = 1, 2,.., N. Deal, Sethi, and Thompson (1979) formulated and solved an advertising game with two players and obtained the open-loop Nash solution by solving a two-point boundary value problem. In Exercise 12.1, you are asked to formulate their problem. See also Deal (1979). Closed-Loop Nash Solution A closed-loop Nash solution is defined when (12.12) is defined in terms of the state of the system. To avoid confusion, we let u i (x, t) = φ i (x, t), i = 1, 2,..., N. (12.17) For these controls to represent a Nash strategy, we must recognize the dependence of the other players actions on the state variable x. Therefore, we need to replace the adjoint equation (12.15) by λ i N N = Hx i Hu i j φ j x = Hx i Hu i φ j x. (12.18) j j=1 j=1,j i The presence of the summation term in (12.18) makes the necessary condition for the closed-loop solution virtually useless for deriving computational algorithms; see Starr and Ho (1969). It is, however, possible to use a dynamic programming approach for solving extremely simple nonzero-sum games, which require the solution of a partial differential equation. In Exercise 12.2, you are asked to formulate this partial differential equation for N = 2. Note that the troublesome summation term in (12.18) is absent in three important cases: (a) in optimal control problems (N = 1) since H u u x =, (b) in two-person zero-sum games because H 1 = H 2 so that H 1 u 2 u 2 x = H 2 u 2 u 2 x = and H 2 u 1 u 1 x = H 1 u 1 u 1 x =, and (c) in openloop nonzero-sum games because u j x =. It certainly is to be expected, therefore, that the closed-loop and open-loop Nash solutions are going to be different, in general. This can be shown explicitly for the linearquadratic case. We conclude this section by providing an interpretation to the adjoint variable λ i. It is the sensitivity of ith player s profits to a perturbation in the state vector. If the other players are using closed-loop (i.e., feedback) strategies, any perturbation δx in the state vector causes them to revise their controls by the amount φ j xδx. If the ith Hamiltonian H i were extremized with respect to u j, j i, this would not affect the
5 Differential Games, Distributed Systems, and Impulse Control ith player s profit; but since H i / u j for i j, the reactions of the other players to the perturbation influence the ith player s profit, and the ith player must account for this effect in considering variations of the trajectory An Application to the Common-Property Fishery Resources Consider extending the fishery model of Section 1.1 by assuming that there are two producers having unrestricted rights to exploit the fish stock in competition with each other. This gives rise to a nonzero-sum differential game analyzed by Clark (1976). Equation (1.2) is modified by ẋ = g(x) q 1 u 1 x q 2 u 2 x, x() = x, (12.19) where u i (t) represents the rate of fishing effort and q i u i x is the rate of catch for the ith producer, i = 1, 2. The control constraints are the state constraints are u i (t) U i, i = 1, 2, (12.2) x(t), (12.21) and the objective function for the ith producer is the total present value of his profits, namely, J i = (p i q i x c i )u i e ρt dt, i = 1, 2. (12.22) To find the Nash solution for this model, we let x i denote the turnpike (or optimal biomass) level given by (1.12) on the assumption that the ith producer is the sole-owner of the fishery. Let the bionomic equilibrium x i b for producer i be defined by (1.4), i.e., x i b = ci p i qi. (12.23) As shown in Exercise 1.2, x i b < xi. If the other producer is not fishing, then producer i can maintain x i b by making the fishing effort u i b = g(xi b )pi c i ; (12.24)
6 12.1. Differential Games 313 here we have assumed U i to be sufficiently large so that u i b Ui. We also assume that x 1 b < x 2 b, (12.25) which means that producer 1 is more efficient than producer 2, i.e., producer 1 can make a positive profit at any level in the interval (x 1 b, x2 b ], while producer 2 loses money in the same interval, except at x 2 b, where he breaks even. For x > x 2 b, both producers make positive profits. Since U 1 u 1 b by assumption, producer 1 has the capability of driving the fish stock to a level down to at least x 1 b which, by (12.25), is less than x 2 b. This implies that producer 2 cannot operate at a sustained level above x 2 b ; and at a sustained level below x2 b, he cannot make a profit. Hence, his optimal policy is bang-bang: U 2 if x > x 2 u 2 b (x) =, (12.26) if x x 2 b. As far as producer 1 is concerned, he wants to attain his turnpike level x 1 if x 1 x 2 b. If x1 > x 2 b and if x x 1, then from (12.26) producer 2 will fish at his maximum rate until the fish stock is driven to x 2 b. At this level it is optimal for producer 1 to fish at a rate which maintains the fish stock at level x 2 b in order to keep producer 2 from fishing. Thus, the optimal policy for producer 1 can be stated as U 1 if x > x 1 u 1 (x) = ū 1 = g( x1 ) q 1 x if x = x 1, if x 1 < x 2 b, (12.27) 1 if x < x 1 u 1 (x) = U 1 g(x 2 b ) q 1 x 2 b if x > x 2 b if x = x 2 b if x < x 2 b, if x 1 x 2 b. (12.28) The formal proof that policies (12.26)-(12.28) give a Nash solution requires direct verification using the result of Section The Nash solution for this case means that for all feasible paths u 1 and u 2, J 1 (u 1, u 2 ) J 1 (u 1, u 2 ), (12.29)
7 Differential Games, Distributed Systems, and Impulse Control and J 2 (u 1, u 2 ) J 2 (u 1, u 2 ). (12.3) The direct verification involves defining a modified growth function g(x) q 2 U 2 x if x > x 2 g 1 b (x) =, g(x) if x x 2 b, and using the Green s theorem results of Section Since U 2 u 2 b by assumption, we have g 1 (x) for x x 2 b. From (1.12) with g replaced by g 1, it can be shown that the new turnpike level for producer 1 is min( x 1, x 2 b ), which defines the optimal policy (12.27)-(12.28) for producer 1. The optimality of (12.26) for producer 2 follows easily. To interpret the results of the model, suppose that producer 1 originally has sole possession of the fishery, but anticipates a rival entry. Producer 1 will switch from his own optimal sustained yield ū 1 to a more intensive exploitation policy prior to the anticipated entry. We can now guess the results in situations involving N producers. The fishery will see the progressive elimination of inefficient producers as the stock of fish decreases. Only the most efficient producers will survive. If, ultimately, two or more maximally efficient producers exist, the fishery will converge to a classical bionomic equilibrium, with zero sustained economic rent. We have now seen that a Nash competitive solution involving N 2 producers results in the long-run dissipation of economic rents. This conclusion depends on the assumption that producers face an infinitely elastic supply of all factors of production going into the fishing effort, but typically the methods of licensing entrants to regulated fisheries make some attempt also to control the factors of production such as permitting the licensee to operate only a single vessel of specific size. In order to develop a model for licensing of fishermen, we let the control variable v i denote the capital stock of the ith producer and let the concave function f(v i ), with f() =, denote the fishing mortality function, for i = 1, 2,..., N. This requires the replacement of q i u i in the previous model by f(v i ). The extended model becomes nonlinear in control variables. You are asked in Exercise 12.2 to formulate this new model and develop necessary conditions for a closed-loop Nash solution for this model with N producers. The reader is referred to Clark (1976) for further details.
8 12.2. Distributed Parameter Systems 317 t and let S(y) be the value of one unit of x(t, y) at time T. Then the objective function is: max u(t,y) Ω T + { J = T h Q(t)x(t, h)dt + F(t, y, x(t, y), u(t, y))dydt } h where Ω is the set of allowable controls. S(y)x(T, y)dy, (12.35) The Distributed Parameter Maximum Principle We will formulate, without giving proofs, a procedure for solving the problem in (12.31)-(12.35) by a distributed parameter maximum principle, which is analogous to the ordinary one. A more complete treatment of this maximum principle can be found in Sage (1968, Chapter 7), Butkowskiy (1969), Lions (1971), Derzko, Sethi, and Thompson (1984), and Haurie, Sethi, and Hartl (1984). In order to obtain necessary conditions for a maximum, we introduce the Hamiltonian H = F + λf, (12.36) where the spatial adjoint function λ(t, y) satisfies λ t = H x + t [ H x t ] + [ ] H, (12.37) y x y where x t = x/ t and x y = x/ y. The boundary conditions on λ are stated for the Γ 2 part of the boundary of D (see Figure 12.1) as follows: λ(t, h) = Q(t) (12.38) and λ(t, y) = S(y). (12.39) Once again we need a consistency requirement similar to (12.34). It is λ(t, h) = Q(T) = S(h), (12.4) which gives the consistency requirement in the sense that the price and the salvage value of a unit x(t, h) must agree.
9 Differential Games, Distributed Systems, and Impulse Control We let u (t, y) denote the optimal control function. Then the distributed parameter maximum principle requires that H(t, y, x, x t,x y, u, λ) H(t, y, x, x t,x y, u, λ) (12.41) for all (t, y) D and all u Ω. We have stated only a simple form of the distributed parameter maximum principle which is sufficient for the cattle ranching example dealt with in the next section. More general forms of the maximum principle are available in the references cited earlier. Among other things, these general forms allow for the function F in (12.35) to contain arguments such as x/ y, 2 x/ y 2, etc. It is also possible to consider controls on the boundary. In this case v(t) in (12.33) will become a control variable The Cattle Ranching Problem Let t denote time and y denote the age of an animal. Let x(t, y) denote the number of cattle of age y on the ranch at time t. Let h be the age at maturity at which the cattle are slaughtered. Thus, the set [, h] is the set of all possible ages of the cattle. Let u(t, y) be the rate at which y-aged cattle are bought at time t, where we agree that a negative value of u denotes a sale. To develop the dynamics of the process it is easy to see that x(t + t, y) = x(t, y t) + u(t, y) t. (12.42) Subtracting x(t, y) from both sides of (12.42), dividing by t, and taking the limit as t, yields the state equation x t = x + u. (12.43) y The boundary and consistency conditions for x are given in (12.32)- (12.34). Here x (y) denotes the initial distribution of cattle at various ages, and v(t) is an exogenously specified breeding rate. To develop the objective function for the cattle rancher, we let T denote the horizon time. Let P(t, y) be the purchase or sale price of a y-aged animal at time t. Let P(t, h) = Q(t) be the slaughter value at time t and let P(T, y) = S(y) be the salvage value of a y-aged animal at the horizon time T. The functions Q and S represent the proceeds of the cattle ranching business. To obtain the profit function we must subtract the costs of running the ranch from these proceeds. Let C(y)
10 12.3. Impulse Control 325 The objective function of the oil driller is to maximize his profit, which is the difference between the revenue from the oil pumped minus the cost of drilling new wells. Since the cost of pumping the oil is fixed, we ignore it. The objective is to T maximize J = N(t) Pbx(t)dt i=1 Qv(t i ), (12.55) where P is the unit price of oil and Q is the cost of drilling a well having an initial stock of 1. Therefore, P is the total value of the entire stock of oil in a new well. See Case (1979) for a re-interpretation of this problem as The Optimal Painting of a Roadside Inn. In the next section we will discuss a maximum principle for impulse optimal control problems, based on the work by Blaquière (1979). The solution of the oil driller s problem appears in Section The Maximum Principle for Impulse Optimal Control In (2.4) we stated the basic optimal control problem with state variable x and an ordinary control variable u. Now we will add an impulse control variable v Ω v and two associated functions. The first function is G(x, v, t), which represents the profit associated with the impulse control. The second function is g(x, v, t), which represents the instantaneous finite change in the state variable when the impulse control is applied. With this notation we can state the impulse optimal control problem as: T max J = subject to N(T) F(x, u, t)dt + i=1 G[x(t i ), v(t i ), t i ] + S[x(T)] N(T) ẋ(t) = f(x(t), u(t), t) + g[x(t i ), v(t i ), t i ], x() = x, u Ω u, v Ω v. As before we will have i=1 (12.56) x(t + i ) = x(t i) + g[x(t i ), v(t i ), t i ] (12.57)
11 Differential Games, Distributed Systems, and Impulse Control at the times t i at which impulse control is applied, i.e., g[x(t i ), v(t i ), t i ] >. Blaquière (1978) has developed the maximum principle necessary optimality conditions to deal with the problem in (12.56). To state these conditions we first define the ordinary Hamiltonian function H(x, u, λ, t) = F(x, u, t) + λf(x, u, t) (12.58) and the impulse Hamiltonian function H I (x(t), v(t), λ(t + ), t) = G(x(t), v(t), t) + λ(t + )g(x(t), v(t), t). (12.59) We now state the impulse maximum principle. Let u, v (t i ), i = 1,..., N (T), where t i > t i 1 >, be optimal controls with the associated x representing an optimal trajectory for (12.56). Then there exists an adjoint variable λ such that the following conditions hold: (i) ẋ = f(x, u, t), t [, T],, t t i, (ii) x (t + i ) = x (t i ) + g[x (t i ), v (t i ), t i ], (iii) λ = Hx [x, u, λ, t], λ(t) = S x [x (T)], t t i, (iv) λ(t i ) = λ(t + i ) + HI x[x (t i ), v (t i ), t i ], (v) H[x, u, λ, t] H[x, u, λ, t] for all u Ω u, t t i, (vi) H I [x (t i ), v (t i ),λ(t + i ), t i] H I [x (t i ), v,λ(t + i ), t i] v Ω v, (vii) H[x (t + i ), u (t + i ), λ(t+ i ), t i] + H I t [x (t + i ), v (t + i ), λ(t+ i ), t i] = H[x (t i ), u (t i ), λ(t i ), t i ] + H I t [x (t i ), v (t i ), λ(t + i ), t i]. (12.6) If t 1 = then the equality sign in (vii) should be replaced by a sign when i = 1. Clearly (i) and (ii) are equivalent to the state equation in (12.56) with the optimum value substituted for the variables. Note that condition (vii) involves the partial derivative of H I with respect to t. Thus, in autonomous problems, where H I t =, condition (vii) means that the Hamiltonian H is continuous at those times where an impulse control is applied.
12 12.3. Impulse Control Solution of the Oil Driller s Problem We now give a solution to the oil driller s problem under the assumption that T is sufficiently small so that no more than one drilling will be found to be optimal. We restate below the problem in Section for this case when t 1 is the drilling time to be determined: { max J = subject to T } Pbx(t)dt Qv(t 1 ) ẋ(t) = bx(t) + δ(t t 1 )v(t)[1 x(t)], x() = 1, v(t) 1. (12.61) To apply the maximum principle we define the Hamiltonian functions corresponding to (12.58) and (12.59): and H(x, λ) = Pbx + λ( bx) = bx(p λ) (12.62) H I (x(t), v(t), λ(t + )) = Qv(t) + λ(t + )v(t)(1 x(t)). (12.63) The ordinary Hamiltonian (12.62) is without an ordinary control variable because we do not have any ordinary control variable in this problem. We now apply the necessary conditions of (12.6) to the oil driller s problem: ẋ = bx, t [, T], t t 1, (12.64) x[t + 1 ] = x(t 1) + v(t 1 )[1 x(t 1 )], (12.65) λ = b(p λ), λ(t) =, t t 1, (12.66) λ(t 1 ) = λ(t + 1 ) v(t 1)λ(t + 1 ), (12.67) [ Q+λ(t + 1 )(1 x)]v (t i ) [ Q+λ(t + 1 )(1 x)]v for v [, 1], (12.68) bx(t + 1 )[P λ(t+ 1 )] = bx(t 1)[P λ(t 1 )]. (12.69) The solution of (12.66) for t > t 1, where t 1 is the drilling time, is λ(t) = P[1 e b(t t) ], t (t 1, T]. (12.7)
13 Differential Games, Distributed Systems, and Impulse Control From (12.68), the optimal impulse control at t 1 is v (t 1 ) = bang[, 1; Q + λ(t + 1 ){1 x(t 1)}]. (12.71) Note that the optimal impulse control is bang-bang, because the impulse Hamiltonian H I in (12.63) is linear in v. The region in the (t, x)-space in which there is certain to be no drilling is given by the set of all (t, x) for which Q + λ(t + )(1 x) <. (12.72) After the drilling, λ(t + 1 ) = P[1 e b(t t 1) ], which we obtain by using (12.7). Substituting this value into (12.72) gives the condition Q + P[1 e b(t t 1) ](1 x) <. (12.73) The boundary of the no-drilling region is obtained when the < sign in (12.73) is replaced by an = sign. This yields the curve x(t 1 ) = [ ] Q 1 P(1 e b(t t1) := φ(t 1 ), (12.74) ) shown in Figure Note that we did not label the region below the curve represented by the complement of (12.73) with v = 1, since (12.73) uses the value of λ(t), t > t 1, and its complement, therefore, does not Figure 12.4: Boundary of No-Drilling Region
14 12.3. Impulse Control 329 represent the condition Q + λ(t + )(1 x) > prior to drilling. Figure 12.4 is drawn under the assumption that φ() = 1 Q P[1 e bt >, (12.75) ] so that the solution is nontrivial; see Exercises 12.5 and In order for (12.75) to hold, it is clear that Q must be relatively small, and P, b, and/or T must be relatively large. In the contrary situation, φ() holds and by Exercise 12.6, no drilling is optimal. The bang-bang nature of the optimal impulse control in (12.71) means that we have v (t 1 ) = 1. Using this in (12.65) and (12.67), we have From (12.7), x(t + 1 ) = 1 and λ(t 1) =. (12.76) λ(t + 1 ) = P[1 eb(t t 1) ]. Substituting these results into (12.69) and solving we get x(t 1 ) = e b(t t 1) : = ψ(t 1 ). (12.77) The curve x = ψ(t) = e b(t t) is drawn in Figure The part BC of the ψ curve lies in the no drilling region, which is above the φ Figure 12.5: Drilling Time
15 Differential Games, Distributed Systems, and Impulse Control With this notation, the impulse optimal control problem is: { } T max J = (πx u)dt + v(t 1 )[Kx(t 1 ) C] subject to ẋ = bx + gu + δ(t t 1 )v(t)[1 x(t)], x() = 1, (12.79) u U, v 1. In writing (12.79) we have assumed that a fraction θ of a machine with quality x has a quality θx. Furthermore, we note that the solution of the state equation will always satisfy x 1, because of the assumption that u U b/g Application of the Impulse Maximum Principle To apply the maximum principle, we define the Hamiltonian functions: and H(x, u, λ) = πx u + λ( bx + gu) = (π λb)x + ( 1 + λg)u (12.8) H I (x(t), v(t), λ(t + )) = [Kx(t) C + λ(t + 1 )(1 x(t))]v(t), (12.81) where λ is defined below and t 1 is the time at which an impulse control is applied. We can now state the necessary conditions (12.6) of optimality for the machine maintenance and the replacement model: ẋ = bx + gu, t t 1, x() = 1, (12.82) x(t + 1 ) = x(t 1) + v(t 1 )[1 x(t 1 )], (12.83) λ = π + λb, λ(t) =, t t 1, (12.84) λ(t 1 ) = λ(t + 1 ) + [K λ(t+ 1 )]v(t 1), (12.85) (π λb)x + ( 1 + λg)u (π λb)x + ( 1 + λg)u, u [, U], {Kx(t 1 ) C + λ(t + 1 )[1 x(t 1)]}v (t 1 ) {Kx(t 1 ) C + λ(t + 1 )[1 x(t 1)]}v, v [, 1], (12.86) (12.87)
16 Differential Games, Distributed Systems, and Impulse Control From (12.87), the optimal replacement control is also bang-bang: v (t 1 ) = bang[, 1; Kx(t 1 ) C + λ(t + 1 ){1 x(t 1)}]. (12.92) As in Section , we set the argument of (12.92) to zero to obtain the curve φ(t): x(t 1 ) = C λ(t+ 1 ) K λ(t + 1 ) = C π[1 e b(t t1) ]/b K π[1 e b(t t 1) ]/b : = φ(t 1). (12.93) To graph φ(t), we compute and compute the time φ() = C π[1 e bt ]/b K π[1 e bt ]/b, (12.94) ˆt = T + 1 b ln(1 bc π ), (12.95) which makes φ(ˆt) =. For simplicity of analysis, we assume gk 1 < gc < gπ/b and ˆt >. (12.96) In Exercise 12.1, you are asked to show that, < ˆt < t 2 < T. (12.97) We can now graph φ(t) as shown in Figure In plotting Figure 12.8, we have assumed that < φ() < 1. This is certainly the case, if T is not too large so that C < (π/b)(1 e bt ). As in the oil driller s problem, we obtain ψ(t) by using (12.88). From (12.92), we have v (t 1 ) = 1 and, therefore, λ(t 1 ) = K from (12.85) and x(t + 1 ) = 1 from (12.83). Since gk 1 from (12.96), we have 1 + gλ(t 1 ) = 1 + gk and, thus, u (t 1 ) = from (12.9). That is, zero maintenance is optimal on the old machine just before it is replaced. Since t 1 < ˆt < t 2 from (12.97), we have u (t + 1 ) = U from Figure That is, full maintenance is optimal on the new machine at the beginning.
17 12.3. Impulse Control 335 Figure 12.8: Replacement Time t 1 and Maintenance Policy Substituting all these values in (12.88), we have π π b [ ] { 1 e b(t t 1) b π } b [1 e b(t t1) g] U = πx(t 1 ) Kbx(t 1 ), (12.98) which yields x(t 1 ) = U(gπ b) + π(b gu)e b(t t 1) b(π Kb) := ψ(t 1 ). (12.99) The graph of ψ(t) is shown in Figure In this figure, AB represents the replacement curve. The optimal trajectory x (t) is shown by CDEFG under the assumption that t 1 > and t 1 < t 2, where t 2 is the intersection point of curves φ(t) and ψ(t), as shown in Figure Remark Figure 12.9 has been drawn for a choice of the problem parameters such that t 1 = t 2. In this case we have the option of either replacing the machine at t 1 = t 2, or not replacing it as sketched in the figure. In other words, this case is one of indifference between replacing and not replacing the machine. If t 1 is larger than t 2, clearly no machine replacement is needed.
18 Differential Games, Distributed Systems, and Impulse Control Figure 12.9: The Case t 1 = t 2 To complete the solution of the problem, it is easy to compute λ(t) in the interval [, t 1 ], and then use (12.9) to obtain the optimal maintenance policy for the old machine before replacement. Using λ(t 1 ) = K obtained above and the adjoint equation (12.84), we have λ(t) = Ke b(t t1) + π [ ] 1 e b(t t 1), t [, t 1 ]. (12.1) b Using (12.1) in (12.9), we can get u (t), t [, t 1 ]. More specifically, we can get the switching point t 1 by solving 1 + λg =. Thus, t 1 = t [ ] πg b b ln. (12.11) πg bkg If t 1, then the policy of no maintenance is optimal in the interval [, t 1 ]. If t 1 >, the optimal maintenance policy for the old machine is U, t [, t u 1 (t) = ], (12.12), t [t 1, t 1]. In plotting Figures 12.7 and 12.8, we have assumed that t 1 >. With (12.12) and (12.82), one can obtain an expression for x(t 1 ) in terms of t 1. Equating this expression to ψ(t 1 ) obtained in (12.99), one
19 Exercises for Chapter can write a transcendental equation for t 1. In Exercise 12.11, you are asked to complete the solution of the problem by obtaining the equation for t 1. EXERCISES FOR CHAPTER A bilinear quadratic advertising model (Deal, Sethi, and Thompson, 1979): Let x i be the market share of firm i and u i be its advertising rate, i = 1, 2. The state equations are ẋ 1 = b 1 u 1 (1 x 1 x 2 ) + e 1 (u 1 u 2 )(x 1 + x 2 ) a 1 x 1 x 1 () = x 1, ẋ 2 = b 2 u 2 (1 x 1 x 2 ) + e 2 (u 2 u 1 )(x 1 + x 2 ) a 2 x 2 x 2 () = x 2, where b i, e i, and a i are given positive constants. Firm i wants to maximize T J i = w i e ρt x i (T) + (c i x i u 2 i )e ρt dt, where w i, c i, and ρ are positive constants. Derive the necessary conditions for the open-loop Nash solution, and formulate the resulting boundary value problem. In a related paper, Deal (1979) provides a numerical solution to this problem with e 1 = e 2 = Develop the nonlinear model described in the last paragraph on page 314 by rewriting (12.19) and (12.22) for the model. Derive the adjoint equation for λ i for the ith producer, and show that the closed-loop Nash policy for producer i is given by f (v i ) = c i (p i λ i )x Verify that (12.48) is a solution of (12.46) Obtain or verify that (12.5) is the solution of (12.43) with initial condition (12.32) and (12.33) Assume φ() > as in (12.75). Show that Q < P and < ˆt < T.
CHAPTER 12 DIFFERENTIAL GAMES, DISTRIBUTED SYSTEMS,
CHAPTER 12 DIFFERENTIAL GAMES, DISTRIBUTED SYSTEMS, AND IMPULSE CONTROL Chapter 12 p. 1/91 DIFFERENTIAL GAMES, DISTRIBUTED SYSTEMS, AND IMPULSE CONTROL Theory of differential games: There may be more than
More informationCHAPTER 9 MAINTENANCE AND REPLACEMENT. Chapter9 p. 1/66
CHAPTER 9 MAINTENANCE AND REPLACEMENT Chapter9 p. 1/66 MAINTENANCE AND REPLACEMENT The problem of determining the lifetime of an asset or an activity simultaneously with its management during that lifetime
More informationCHAPTER 7 APPLICATIONS TO MARKETING. Chapter 7 p. 1/53
CHAPTER 7 APPLICATIONS TO MARKETING Chapter 7 p. 1/53 APPLICATIONS TO MARKETING State Equation: Rate of sales expressed in terms of advertising, which is a control variable Objective: Profit maximization
More informationWe now impose the boundary conditions in (6.11) and (6.12) and solve for a 1 and a 2 as follows: m 1 e 2m 1T m 2 e (m 1+m 2 )T. (6.
158 6. Applications to Production And Inventory For ease of expressing a 1 and a 2, let us define two constants b 1 = I 0 Q(0), (6.13) b 2 = ˆP Q(T) S(T). (6.14) We now impose the boundary conditions in
More informationCHAPTER 2 THE MAXIMUM PRINCIPLE: CONTINUOUS TIME. Chapter2 p. 1/67
CHAPTER 2 THE MAXIMUM PRINCIPLE: CONTINUOUS TIME Chapter2 p. 1/67 THE MAXIMUM PRINCIPLE: CONTINUOUS TIME Main Purpose: Introduce the maximum principle as a necessary condition to be satisfied by any optimal
More informationDeterministic Dynamic Programming
Deterministic Dynamic Programming 1 Value Function Consider the following optimal control problem in Mayer s form: V (t 0, x 0 ) = inf u U J(t 1, x(t 1 )) (1) subject to ẋ(t) = f(t, x(t), u(t)), x(t 0
More informationCHAPTER 3 THE MAXIMUM PRINCIPLE: MIXED INEQUALITY CONSTRAINTS. p. 1/73
CHAPTER 3 THE MAXIMUM PRINCIPLE: MIXED INEQUALITY CONSTRAINTS p. 1/73 THE MAXIMUM PRINCIPLE: MIXED INEQUALITY CONSTRAINTS Mixed Inequality Constraints: Inequality constraints involving control and possibly
More informationF and G have continuous second-order derivatives. Assume Equation (1.1) possesses an equilibrium point (x*,y*) so that
3.1.6. Characterizing solutions (continued) C4. Stability analysis for nonlinear dynamical systems In many economic problems, the equations describing the evolution of a system are non-linear. The behaviour
More informationMathematical Economics. Lecture Notes (in extracts)
Prof. Dr. Frank Werner Faculty of Mathematics Institute of Mathematical Optimization (IMO) http://math.uni-magdeburg.de/ werner/math-ec-new.html Mathematical Economics Lecture Notes (in extracts) Winter
More informationSolution by the Maximum Principle
292 11. Economic Applications per capita variables so that it is formally similar to the previous model. The introduction of the per capita variables makes it possible to treat the infinite horizon version
More informationHJB equations. Seminar in Stochastic Modelling in Economics and Finance January 10, 2011
Department of Probability and Mathematical Statistics Faculty of Mathematics and Physics, Charles University in Prague petrasek@karlin.mff.cuni.cz Seminar in Stochastic Modelling in Economics and Finance
More informationSolution of Stochastic Optimal Control Problems and Financial Applications
Journal of Mathematical Extension Vol. 11, No. 4, (2017), 27-44 ISSN: 1735-8299 URL: http://www.ijmex.com Solution of Stochastic Optimal Control Problems and Financial Applications 2 Mat B. Kafash 1 Faculty
More informationAn Introduction to Noncooperative Games
An Introduction to Noncooperative Games Alberto Bressan Department of Mathematics, Penn State University Alberto Bressan (Penn State) Noncooperative Games 1 / 29 introduction to non-cooperative games:
More informationChapter Introduction. A. Bensoussan
Chapter 2 EXPLICIT SOLUTIONS OFLINEAR QUADRATIC DIFFERENTIAL GAMES A. Bensoussan International Center for Decision and Risk Analysis University of Texas at Dallas School of Management, P.O. Box 830688
More informationTheory and Applications of Constrained Optimal Control Proble
Theory and Applications of Constrained Optimal Control Problems with Delays PART 1 : Mixed Control State Constraints Helmut Maurer 1, Laurenz Göllmann 2 1 Institut für Numerische und Angewandte Mathematik,
More informationFunctions. A function is a rule that gives exactly one output number to each input number.
Functions A function is a rule that gives exactly one output number to each input number. Why it is important to us? The set of all input numbers to which the rule applies is called the domain of the function.
More informationA Note on Open Loop Nash Equilibrium in Linear-State Differential Games
Applied Mathematical Sciences, vol. 8, 2014, no. 145, 7239-7248 HIKARI Ltd, www.m-hikari.com http://dx.doi.org/10.12988/ams.2014.49746 A Note on Open Loop Nash Equilibrium in Linear-State Differential
More informationSubject: Optimal Control Assignment-1 (Related to Lecture notes 1-10)
Subject: Optimal Control Assignment- (Related to Lecture notes -). Design a oil mug, shown in fig., to hold as much oil possible. The height and radius of the mug should not be more than 6cm. The mug must
More informationContents. Set Theory. Functions and its Applications CHAPTER 1 CHAPTER 2. Preface... (v)
(vii) Preface... (v) CHAPTER 1 Set Theory Definition of Set... 1 Roster, Tabular or Enumeration Form... 1 Set builder Form... 2 Union of Set... 5 Intersection of Sets... 9 Distributive Laws of Unions and
More informationproblem. max Both k (0) and h (0) are given at time 0. (a) Write down the Hamilton-Jacobi-Bellman (HJB) Equation in the dynamic programming
1. Endogenous Growth with Human Capital Consider the following endogenous growth model with both physical capital (k (t)) and human capital (h (t)) in continuous time. The representative household solves
More informationAlberto Bressan. Department of Mathematics, Penn State University
Non-cooperative Differential Games A Homotopy Approach Alberto Bressan Department of Mathematics, Penn State University 1 Differential Games d dt x(t) = G(x(t), u 1(t), u 2 (t)), x(0) = y, u i (t) U i
More informationEconomic Growth (Continued) The Ramsey-Cass-Koopmans Model. 1 Literature. Ramsey (1928) Cass (1965) and Koopmans (1965) 2 Households (Preferences)
III C Economic Growth (Continued) The Ramsey-Cass-Koopmans Model 1 Literature Ramsey (1928) Cass (1965) and Koopmans (1965) 2 Households (Preferences) Population growth: L(0) = 1, L(t) = e nt (n > 0 is
More informationHomework Solution # 3
ECSE 644 Optimal Control Feb, 4 Due: Feb 17, 4 (Tuesday) Homework Solution # 3 1 (5%) Consider the discrete nonlinear control system in Homework # For the optimal control and trajectory that you have found
More informationMathematical Foundations -1- Constrained Optimization. Constrained Optimization. An intuitive approach 2. First Order Conditions (FOC) 7
Mathematical Foundations -- Constrained Optimization Constrained Optimization An intuitive approach First Order Conditions (FOC) 7 Constraint qualifications 9 Formal statement of the FOC for a maximum
More information2 Functions and Their
CHAPTER Functions and Their Applications Chapter Outline Introduction The Concept of a Function Types of Functions Roots (Zeros) of a Function Some Useful Functions in Business and Economics Equilibrium
More informationminimize x subject to (x 2)(x 4) u,
Math 6366/6367: Optimization and Variational Methods Sample Preliminary Exam Questions 1. Suppose that f : [, L] R is a C 2 -function with f () on (, L) and that you have explicit formulae for
More informationHandout 1: Introduction to Dynamic Programming. 1 Dynamic Programming: Introduction and Examples
SEEM 3470: Dynamic Optimization and Applications 2013 14 Second Term Handout 1: Introduction to Dynamic Programming Instructor: Shiqian Ma January 6, 2014 Suggested Reading: Sections 1.1 1.5 of Chapter
More informationOptimal Control. Macroeconomics II SMU. Ömer Özak (SMU) Economic Growth Macroeconomics II 1 / 112
Optimal Control Ömer Özak SMU Macroeconomics II Ömer Özak (SMU) Economic Growth Macroeconomics II 1 / 112 Review of the Theory of Optimal Control Section 1 Review of the Theory of Optimal Control Ömer
More informationImpulsive Stabilization and Application to a Population Growth Model*
Nonlinear Dynamics and Systems Theory, 2(2) (2002) 173 184 Impulsive Stabilization and Application to a Population Growth Model* Xinzhi Liu 1 and Xuemin Shen 2 1 Department of Applied Mathematics, University
More informationContents CONTENTS 1. 1 Straight Lines and Linear Equations 1. 2 Systems of Equations 6. 3 Applications to Business Analysis 11.
CONTENTS 1 Contents 1 Straight Lines and Linear Equations 1 2 Systems of Equations 6 3 Applications to Business Analysis 11 4 Functions 16 5 Quadratic Functions and Parabolas 21 6 More Simple Functions
More informationSCHOOL OF DISTANCE EDUCATION
SCHOOL OF DISTANCE EDUCATION CCSS UG PROGRAMME MATHEMATICS (OPEN COURSE) (For students not having Mathematics as Core Course) MM5D03: MATHEMATICS FOR SOCIAL SCIENCES FIFTH SEMESTER STUDY NOTES Prepared
More informationMathematics for Economics ECON MA/MSSc in Economics-2017/2018. Dr. W. M. Semasinghe Senior Lecturer Department of Economics
Mathematics for Economics ECON 53035 MA/MSSc in Economics-2017/2018 Dr. W. M. Semasinghe Senior Lecturer Department of Economics MATHEMATICS AND STATISTICS LERNING OUTCOMES: By the end of this course unit
More informationAdvanced Microeconomics
Advanced Microeconomics Leonardo Felli EC441: Room D.106, Z.332, D.109 Lecture 8 bis: 24 November 2004 Monopoly Consider now the pricing behavior of a profit maximizing monopolist: a firm that is the only
More informationCalculus in Business. By Frederic A. Palmliden December 7, 1999
Calculus in Business By Frederic A. Palmliden December 7, 999 Optimization Linear Programming Game Theory Optimization The quest for the best Definition of goal equilibrium: The equilibrium state is defined
More informationProject Optimal Control: Optimal Control with State Variable Inequality Constraints (SVIC)
Project Optimal Control: Optimal Control with State Variable Inequality Constraints (SVIC) Mahdi Ghazaei, Meike Stemmann Automatic Control LTH Lund University Problem Formulation Find the maximum of the
More informationOptimal control theory with applications to resource and environmental economics
Optimal control theory with applications to resource and environmental economics Michael Hoel, August 10, 2015 (Preliminary and incomplete) 1 Introduction This note gives a brief, non-rigorous sketch of
More informationDuality Theory of Constrained Optimization
Duality Theory of Constrained Optimization Robert M. Freund April, 2014 c 2014 Massachusetts Institute of Technology. All rights reserved. 1 2 1 The Practical Importance of Duality Duality is pervasive
More information(a) Write down the Hamilton-Jacobi-Bellman (HJB) Equation in the dynamic programming
1. Government Purchases and Endogenous Growth Consider the following endogenous growth model with government purchases (G) in continuous time. Government purchases enhance production, and the production
More informationNew Notes on the Solow Growth Model
New Notes on the Solow Growth Model Roberto Chang September 2009 1 The Model The firstingredientofadynamicmodelisthedescriptionofthetimehorizon. In the original Solow model, time is continuous and the
More informationEndogenous Growth Theory
Endogenous Growth Theory Lecture Notes for the winter term 2010/2011 Ingrid Ott Tim Deeken October 21st, 2010 CHAIR IN ECONOMIC POLICY KIT University of the State of Baden-Wuerttemberg and National Laboratory
More informationThe Envelope Theorem
The Envelope Theorem In an optimization problem we often want to know how the value of the objective function will change if one or more of the parameter values changes. Let s consider a simple example:
More informationHARVESTING IN A TWO-PREY ONE-PREDATOR FISHERY: A BIOECONOMIC MODEL
ANZIAM J. 452004), 443 456 HARVESTING IN A TWO-PREY ONE-PREDATOR FISHERY: A BIOECONOMIC MODEL T. K. KAR 1 and K. S. CHAUDHURI 2 Received 22 June, 2001; revised 20 September, 2002) Abstract A multispecies
More informationLinear Quadratic Zero-Sum Two-Person Differential Games Pierre Bernhard June 15, 2013
Linear Quadratic Zero-Sum Two-Person Differential Games Pierre Bernhard June 15, 2013 Abstract As in optimal control theory, linear quadratic (LQ) differential games (DG) can be solved, even in high dimension,
More informationIntroduction to optimal control theory in continuos time (with economic applications) Salvatore Federico
Introduction to optimal control theory in continuos time (with economic applications) Salvatore Federico June 26, 2017 2 Contents 1 Introduction to optimal control problems in continuous time 5 1.1 From
More informationChapter 5. Pontryagin s Minimum Principle (Constrained OCP)
Chapter 5 Pontryagin s Minimum Principle (Constrained OCP) 1 Pontryagin s Minimum Principle Plant: (5-1) u () t U PI: (5-2) Boundary condition: The goal is to find Optimal Control. 2 Pontryagin s Minimum
More information5.3. Exercises on the curve analysis of polynomial functions
.. Exercises on the curve analysis of polynomial functions Exercise : Curve analysis Examine the following functions on symmetry, x- and y-intercepts, extrema and inflexion points. Draw their graphs including
More informationDynamic stochastic game and macroeconomic equilibrium
Dynamic stochastic game and macroeconomic equilibrium Tianxiao Zheng SAIF 1. Introduction We have studied single agent problems. However, macro-economy consists of a large number of agents including individuals/households,
More informationStochastic Equilibrium Problems arising in the energy industry
Stochastic Equilibrium Problems arising in the energy industry Claudia Sagastizábal (visiting researcher IMPA) mailto:sagastiz@impa.br http://www.impa.br/~sagastiz ENEC workshop, IPAM, Los Angeles, January
More informationHyperbolicity of Systems Describing Value Functions in Differential Games which Model Duopoly Problems. Joanna Zwierzchowska
Decision Making in Manufacturing and Services Vol. 9 05 No. pp. 89 00 Hyperbolicity of Systems Describing Value Functions in Differential Games which Model Duopoly Problems Joanna Zwierzchowska Abstract.
More informationPart A: Answer question A1 (required), plus either question A2 or A3.
Ph.D. Core Exam -- Macroeconomics 5 January 2015 -- 8:00 am to 3:00 pm Part A: Answer question A1 (required), plus either question A2 or A3. A1 (required): Ending Quantitative Easing Now that the U.S.
More informationAIMS Exercise Set # 1
AIMS Exercise Set #. Determine the form of the single precision floating point arithmetic used in the computers at AIMS. What is the largest number that can be accurately represented? What is the smallest
More informationRobust control and applications in economic theory
Robust control and applications in economic theory In honour of Professor Emeritus Grigoris Kalogeropoulos on the occasion of his retirement A. N. Yannacopoulos Department of Statistics AUEB 24 May 2013
More informationSF2972 Game Theory Exam with Solutions March 15, 2013
SF2972 Game Theory Exam with s March 5, 203 Part A Classical Game Theory Jörgen Weibull and Mark Voorneveld. (a) What are N, S and u in the definition of a finite normal-form (or, equivalently, strategic-form)
More informationRamsey Cass Koopmans Model (1): Setup of the Model and Competitive Equilibrium Path
Ramsey Cass Koopmans Model (1): Setup of the Model and Competitive Equilibrium Path Ryoji Ohdoi Dept. of Industrial Engineering and Economics, Tokyo Tech This lecture note is mainly based on Ch. 8 of Acemoglu
More informationDYNAMIC LECTURE 5: DISCRETE TIME INTERTEMPORAL OPTIMIZATION
DYNAMIC LECTURE 5: DISCRETE TIME INTERTEMPORAL OPTIMIZATION UNIVERSITY OF MARYLAND: ECON 600. Alternative Methods of Discrete Time Intertemporal Optimization We will start by solving a discrete time intertemporal
More informationEE291E/ME 290Q Lecture Notes 8. Optimal Control and Dynamic Games
EE291E/ME 290Q Lecture Notes 8. Optimal Control and Dynamic Games S. S. Sastry REVISED March 29th There exist two main approaches to optimal control and dynamic games: 1. via the Calculus of Variations
More informationEN Applied Optimal Control Lecture 8: Dynamic Programming October 10, 2018
EN530.603 Applied Optimal Control Lecture 8: Dynamic Programming October 0, 08 Lecturer: Marin Kobilarov Dynamic Programming (DP) is conerned with the computation of an optimal policy, i.e. an optimal
More informationPh.D. Preliminary Examination MICROECONOMIC THEORY Applied Economics Graduate Program May 2012
Ph.D. Preliminary Examination MICROECONOMIC THEORY Applied Economics Graduate Program May 2012 The time limit for this exam is 4 hours. It has four sections. Each section includes two questions. You are
More informationOPTIMAL CONTROL CHAPTER INTRODUCTION
CHAPTER 3 OPTIMAL CONTROL What is now proved was once only imagined. William Blake. 3.1 INTRODUCTION After more than three hundred years of evolution, optimal control theory has been formulated as an extension
More informationPrinciples of Optimal Control Spring 2008
MIT OpenCourseWare http://ocw.mit.edu 16.323 Principles of Optimal Control Spring 2008 For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms. 16.323 Lecture
More informationGROWTH IN LENGTH: a model for the growth of an individual
GROWTH IN LENGTH: a model for the growth of an individual Consider the following differential equation: L' L What are the dimensions for parameters and? As was the case with the differential equation relating
More informationNoncooperative continuous-time Markov games
Morfismos, Vol. 9, No. 1, 2005, pp. 39 54 Noncooperative continuous-time Markov games Héctor Jasso-Fuentes Abstract This work concerns noncooperative continuous-time Markov games with Polish state and
More informationMathematical Appendix. Ramsey Pricing
Mathematical Appendix Ramsey Pricing PROOF OF THEOREM : I maximize social welfare V subject to π > K. The Lagrangian is V + κπ K the associated first-order conditions are that for each I + κ P I C I cn
More informationDifferential Games II. Marc Quincampoix Université de Bretagne Occidentale ( Brest-France) SADCO, London, September 2011
Differential Games II Marc Quincampoix Université de Bretagne Occidentale ( Brest-France) SADCO, London, September 2011 Contents 1. I Introduction: A Pursuit Game and Isaacs Theory 2. II Strategies 3.
More informationIntroduction to Continuous-Time Dynamic Optimization: Optimal Control Theory
Econ 85/Chatterjee Introduction to Continuous-ime Dynamic Optimization: Optimal Control heory 1 States and Controls he concept of a state in mathematical modeling typically refers to a specification of
More informationMathematical Economics: Lecture 16
Mathematical Economics: Lecture 16 Yu Ren WISE, Xiamen University November 26, 2012 Outline 1 Chapter 21: Concave and Quasiconcave Functions New Section Chapter 21: Concave and Quasiconcave Functions Concave
More informationLinear Quadratic Zero-Sum Two-Person Differential Games
Linear Quadratic Zero-Sum Two-Person Differential Games Pierre Bernhard To cite this version: Pierre Bernhard. Linear Quadratic Zero-Sum Two-Person Differential Games. Encyclopaedia of Systems and Control,
More informationDynamical Systems. August 13, 2013
Dynamical Systems Joshua Wilde, revised by Isabel Tecu, Takeshi Suzuki and María José Boccardi August 13, 2013 Dynamical Systems are systems, described by one or more equations, that evolve over time.
More informationEconS 501 Final Exam - December 10th, 2018
EconS 501 Final Exam - December 10th, 018 Show all your work clearly and make sure you justify all your answers. NAME 1. Consider the market for smart pencil in which only one firm (Superapiz) enjoys a
More informationAnswer Key: Problem Set 3
Answer Key: Problem Set Econ 409 018 Fall Question 1 a This is a standard monopoly problem; using MR = a 4Q, let MR = MC and solve: Q M = a c 4, P M = a + c, πm = (a c) 8 The Lerner index is then L M P
More informationTutorial on Control and State Constrained Optimal Control Pro. Control Problems and Applications Part 3 : Pure State Constraints
Tutorial on Control and State Constrained Optimal Control Problems and Applications Part 3 : Pure State Constraints University of Münster Institute of Computational and Applied Mathematics SADCO Summer
More informationReflected Brownian Motion
Chapter 6 Reflected Brownian Motion Often we encounter Diffusions in regions with boundary. If the process can reach the boundary from the interior in finite time with positive probability we need to decide
More informationNumerical illustration
A umerical illustration Inverse demand is P q, t = a 0 a 1 e λ 2t bq, states of the world are distributed according to f t = λ 1 e λ 1t, and rationing is anticipated and proportional. a 0, a 1, λ = λ 1
More informationLecture 1. Stochastic Optimization: Introduction. January 8, 2018
Lecture 1 Stochastic Optimization: Introduction January 8, 2018 Optimization Concerned with mininmization/maximization of mathematical functions Often subject to constraints Euler (1707-1783): Nothing
More informationJoint work with Nguyen Hoang (Univ. Concepción, Chile) Padova, Italy, May 2018
EXTENDED EULER-LAGRANGE AND HAMILTONIAN CONDITIONS IN OPTIMAL CONTROL OF SWEEPING PROCESSES WITH CONTROLLED MOVING SETS BORIS MORDUKHOVICH Wayne State University Talk given at the conference Optimization,
More informationMath Ordinary Differential Equations
Math 411 - Ordinary Differential Equations Review Notes - 1 1 - Basic Theory A first order ordinary differential equation has the form x = f(t, x) (11) Here x = dx/dt Given an initial data x(t 0 ) = x
More information16. Working with the Langevin and Fokker-Planck equations
16. Working with the Langevin and Fokker-Planck equations In the preceding Lecture, we have shown that given a Langevin equation (LE), it is possible to write down an equivalent Fokker-Planck equation
More informationDeterministic Optimal Control
Online Appendix A Deterministic Optimal Control As far as the laws of mathematics refer to reality, they are not certain; and as far as they are certain, they do not refer to reality. Albert Einstein (1879
More informationPontryagin s Minimum Principle 1
ECE 680 Fall 2013 Pontryagin s Minimum Principle 1 In this handout, we provide a derivation of the minimum principle of Pontryagin, which is a generalization of the Euler-Lagrange equations that also includes
More informationIncreasingly, economists are asked not just to study or explain or interpret markets, but to design them.
What is market design? Increasingly, economists are asked not just to study or explain or interpret markets, but to design them. This requires different tools and ideas than neoclassical economics, which
More informationOptimization Techniques and Problem Analysis for Managers
Optimization Techniques and Problem Analysis for Managers Optimization is one of the most basic subjects in management and economics. Dynamic programming and Control Problem are powerful tools in related
More informationCalculus. Applications of Differentiations (IV)
Calculus Applications of Differentiations (IV) Outline 1 Rates of Change In Economics And The Sciences Applications of Derivative In Economics Applications of Derivative in the Sciences 2 Related Rate
More informationEconomic Growth: Lecture 9, Neoclassical Endogenous Growth
14.452 Economic Growth: Lecture 9, Neoclassical Endogenous Growth Daron Acemoglu MIT November 28, 2017. Daron Acemoglu (MIT) Economic Growth Lecture 9 November 28, 2017. 1 / 41 First-Generation Models
More informationDeterministic Optimal Control
page A1 Online Appendix A Deterministic Optimal Control As far as the laws of mathematics refer to reality, they are not certain; and as far as they are certain, they do not refer to reality. Albert Einstein
More informationPoint Process Control
Point Process Control The following note is based on Chapters I, II and VII in Brémaud s book Point Processes and Queues (1981). 1 Basic Definitions Consider some probability space (Ω, F, P). A real-valued
More informationSection 1.6 Inverse Functions
0 Chapter 1 Section 1.6 Inverse Functions A fashion designer is travelling to Milan for a fashion show. He asks his assistant, Betty, what 7 degrees Fahrenheit is in Celsius, and after a quick search on
More informationThe Ramsey Model. (Lecture Note, Advanced Macroeconomics, Thomas Steger, SS 2013)
The Ramsey Model (Lecture Note, Advanced Macroeconomics, Thomas Steger, SS 213) 1 Introduction The Ramsey model (or neoclassical growth model) is one of the prototype models in dynamic macroeconomics.
More informationOblivious Equilibrium: A Mean Field Approximation for Large-Scale Dynamic Games
Oblivious Equilibrium: A Mean Field Approximation for Large-Scale Dynamic Games Gabriel Y. Weintraub, Lanier Benkard, and Benjamin Van Roy Stanford University {gweintra,lanierb,bvr}@stanford.edu Abstract
More informationQuadratic function and equations Quadratic function/equations, supply, demand, market equilibrium
Exercises 8 Quadratic function and equations Quadratic function/equations, supply, demand, market equilibrium Objectives - know and understand the relation between a quadratic function and a quadratic
More informationNon-zero Sum Stochastic Differential Games of Fully Coupled Forward-Backward Stochastic Systems
Non-zero Sum Stochastic Differential Games of Fully Coupled Forward-Backward Stochastic Systems Maoning ang 1 Qingxin Meng 1 Yongzheng Sun 2 arxiv:11.236v1 math.oc 12 Oct 21 Abstract In this paper, an
More information1.4 CONCEPT QUESTIONS, page 49
.4 CONCEPT QUESTIONS, page 49. The intersection must lie in the first quadrant because only the parts of the demand and supply curves in the first quadrant are of interest.. a. The breakeven point P0(
More informationErgodicity and Non-Ergodicity in Economics
Abstract An stochastic system is called ergodic if it tends in probability to a limiting form that is independent of the initial conditions. Breakdown of ergodicity gives rise to path dependence. We illustrate
More informationTutorial on Control and State Constrained Optimal Control Problems
Tutorial on Control and State Constrained Optimal Control Problems To cite this version:. blems. SADCO Summer School 211 - Optimal Control, Sep 211, London, United Kingdom. HAL Id: inria-629518
More informationDISCRETE-TIME DYNAMICS OF AN
Chapter 1 DISCRETE-TIME DYNAMICS OF AN OLIGOPOLY MODEL WITH DIFFERENTIATED GOODS K. Andriopoulos, T. Bountis and S. Dimas * Department of Mathematics, University of Patras, Patras, GR-26500, Greece Abstract
More informationToulouse School of Economics, M2 Macroeconomics 1 Professor Franck Portier. Exam Solution
Toulouse School of Economics, 2013-2014 M2 Macroeconomics 1 Professor Franck Portier Exam Solution This is a 3 hours exam. Class slides and any handwritten material are allowed. You must write legibly.
More informationEssential Elements of Continuous Time DynamicOptimization
ONE Essential Elements of Continuous Time DynamicOptimization In order to motivate the following introductory material on dynamic optimization problems, it will be advantageous to draw heavily on your
More informationwhere u is the decision-maker s payoff function over her actions and S is the set of her feasible actions.
Seminars on Mathematics for Economics and Finance Topic 3: Optimization - interior optima 1 Session: 11-12 Aug 2015 (Thu/Fri) 10:00am 1:00pm I. Optimization: introduction Decision-makers (e.g. consumers,
More informationAdvertising and Promotion in a Marketing Channel
Applied Mathematical Sciences, Vol. 13, 2019, no. 9, 405-413 HIKARI Ltd, www.m-hikari.com https://doi.org/10.12988/ams.2019.9240 Advertising and Promotion in a Marketing Channel Alessandra Buratto Dipartimento
More informationConcave programming. Concave programming is another special case of the general constrained optimization. subject to g(x) 0
1 Introduction Concave programming Concave programming is another special case of the general constrained optimization problem max f(x) subject to g(x) 0 in which the objective function f is concave and
More informationE-Companion to. Managing Co-Creation in Information Technology Projects: A Differential Games Approach
Article submitted to Information Systems Research E-Companion to Managing Co-Creation in Information Technology Projects: A Differential Games Approach Appendix A: Proofs Proof of Lemma We write the equations
More information