Loading [MathJax]/jax/output/HTML-CSS/jax.js

2020年10月14日 星期三

Solution to Partial Differential Equations: Methods and Applications (Robert McOwen) Section 1.2

Solution to Partial Differential Equations: Methods and Applications (Robert McOwen) Section 1.2

  1. Solve the initial value problem a(u)ux+uy=0 with u(x,0)=h(x), and show the solution becomes singular for some y>0 unless a(h(s)) is a nondecreasing function of s.
  2. SolutionConsider the characteristic curve x=x(t), y=y(t), z=z(t)=u(x(t),y(t)). We have the following ordinary differential equation

    {dxdt=a(z),dydt=1,dzdt=0, with the initial condition {x(0)=x0,y(0)=0,z(0)=u(x(0),y(0))=h(x0).

    We can solve these equations to obtain

    {x(t)=x0+a(h(x0))t,y(t)=t,z(t)=h(x0).

    By z(t)=u(x(t),y(t)), we have

    u(x0+a(h(x0))t,t)=h(x0).

    By taking t=y, x0=xa(u(x,y))y, we can deduce that

    u(x,y)=h(xa(u(x,y))y).

    Now, for x1x0, we consider another characteristic curve x=x(s), y=y(s), z(s)=u(x(s),y(s)). Then, we can obtain

    {x(s)=x1+a(h(x1))s,y(s)=s,z(s)=h(x1).

    Without loss of generality, we may assume that x1>x0.
    • Suppose ah is not a nondecreasing function of. Then there exists s1>s2 such that a(h(s1))<a(h(s2)), which implies h(s1)h(s2). Thus, we choose x0=s2, x1=s1, and

      t=s=y:=s1s2a(h(s2))a(h(s1))>0,

      which implies that (x(t),y(t))=(x(s),y(s)), and hence

      h(s2)=z(t)=u(x(t),y(t))=u(x(s),y(s))=z(s)=h(s1),

      which leads a singularity.
    • Suppose ah is a nondecreasing functinon. When t=s, we can observe that

      x(s)=x1+a(h(x1))s>x0+a(h(x0))t=x(t),

      which means that two characteristic curve cannot have any intersect points unless they are same. Hence the solution has no singularity.

  3. Consider (17) with the intial condition

    u(x,0)=h(x)={u0for x0u0(1x)for 0<x<10for x1,

    where u0>0. Show that a shock develops at a finite time and describe the weak solution.
  4. SolutionIt is well known that the solution has the implicit form u(x,y)=h(xu(x,y)y) and the family of the characteristic curve passing (x,y)=(s,0) is x=h(s)y.
    • If s1, then the characteristic curve is a vertical line x=s.
    • If s0, then the characteristic curve is x=u0y+s whose slope is 1u0.
    • If 0<s<1, then the characteristic curve is x=u0(1s)y+s whose slope is 1u0(1s).
    Note that the characteristic curve passes the point (1,1u0) for all 0<s<1, and hence the solution has the singularity at this point.

    To deal with the singularity, we need to consider the shock wave and introduce the curve

    ξ(y)=G(ur)G(ul)urul=u02

    and the curve passes (1,1u0). Thus, we can get x=ξ(y)=1+u0y2.

    Therefore, we obtain the solution

    u(x,y)={0,if x1,2x1>u0y,u0(1x)1u0y,if 0<x<1, x>u0y,u0,if u0y>x, u0y>2x1.

    The figure of the solution can be shown as below.

  5. Consider (17) with the initial condition

    u(x,0)=h(x)={0for x<0u0(x1)for x>0,

    where u0>0. There is a weak solution u(x,y) that has a jump discontinuity along x=ξ(y). Find this curve and describe the weak solution.
  6. SolutionIt is well known that the solution has the implicit form u(x,y)=h(xu(x,y)y) and the family of the characteristic curve passing (x,y)=(s,0) is x=h(s)y.
    • If s<0, then the characteristic curve is a vertical line x=s.
    • If s>0, then the characteristic curve is a line x=u0(s1)y+s.
      • If s>1, then the slope of the charcteristic curve is 1u0(s1)>0.
      • If s=1, then the characteristic curve is x=1 and u(1,y)=0 for all y>0.
      • If 0<s<1, then the slope of the characteristic curve is 1u0(s1)<0.
    It is clear that u has a singularity at origin. For y>0, we consider the ordinary differential equation

    ξ(y)=G(ur)G(ul)urul=ur(ξ(y),y)2,

    whose value is determined by the characteristic curve passing (ξ(y),y). By taking s=ξ(y)+u0y1+u0y, we obtain

    ξ(y)=u0ξ(y)u02+2u0y,

    which can be written as

    ξ(y)u02+2u0yξ(y)=u02+2u0y.

    Multiplying it by the integral factor (1+u0y)1/2 gives

    ddt(ξ(y)1+u0y)=u02(1+u0y)32,

    which can be solved by

    x=ξ(y)=11+u0y.

    Thus, the solution is

    u(x,y)={0,if x<11+u0y,u0(x1)1+u0y,if x>11+u0y.

    The figure of the solution can be shown as below.

  7. Consider (17) with initial condition

    u(x,0)=h(x)={ulfor x<x0urfor x>x0,

    where ul<ur. Find a continuous weak solution.
  8. SolutionIt is well known that the solution has the implicit form u(x,y)=h(xu(x,y)y) and the family of the characteristic curve passing (x,y)=(s,0) is x=h(s)y.
    • If s>x0, then x=ury+s whose slope is 1ur.
    • If s<x0, then x=uly+s whose slope is 1ul.
    Since ul<ur, the rarefaction wave of the solution is

    u(x,y)={ur,if xx0>ury,xx0y,if uly<xx0<ury,ul,if xx0<uly.

    Moreover, the figure of the solution can be shown as follows.
    • Case 1. 0<ul<ur.
    • Case 2. ul<0<ur.
    • Case 3. ul<ur<0.

  9. Find a simple wave solution u(x,y)=v(x/y) for (19) when G(u)=u4/4. Use this to define a continuous weak solution of (19) for y>0 that satisfies (22) with u0<0.
  10. SolutionAccording to the condition that G(u)=u4/4, the partial differential equation can be rewritten as u3ux+uy=0. Since u(x,y)=v(x/y), we have

    ux(x,y)=v(x/y)y, uy(x,y)=xv(x/y)y2.

    Thus, the partial differential equation can be transformed into v(x/y)=(x/y)1/3, and hence u(x,y)=(x/y)1/3. Now, we consider the characteristic curve x=x(t), y=y(t), z=z(t)=u(x(t),y(t)) satisfying

    {dxdt=z3,dydt=1,dzdt=0 with the initial condition {x(0)=x0,y(0)=0,z(0)=u(x(0),y(0))=h(x0).

    We can solve it by x(t)=h3(x0)t+x0, y(t)=t and z(t)=h(x0). Thus, we obtain

    u(x,y)=h(xu3(x,y)y)

    • If x0>0, then the characteristic curve is vertical line x=x0.
    • If x0<0, then the characteristic curve is the line x=u30y+x0 whose slope is 1u30.
    Hence, the solution can be represented as

    u(x,y)={0,if x>0,(xy)1/3,if u30y<x<0,u0,if x<u30y.

    The figure of the solution can be shown as below.

  11. Consider (17) with (22) where u0<0. In addition to the rarefaction solution described in the text, show that there is a weak solution with a shock along x=u0y/2.
  12. SolutionLet

    u(x,y)={0,if 2x>u0y,u0,if 2x<u0y.

    Fix yR and choose two real a,b satisfying 2a<u0y<2b. We verify directly that u is the weak solution, i.e., to prove the following equality holds:

    G(u(b,y))G(u(a,y))+ddybau(x,y)dx=0.

    A direct computation gives

    G(u(b,y))G(u(a,y))=12u2(b,y)12u2(a,y)=12u20=ddyu0y2au0dx=ddybau(x,y)dx,

    as desired.

  13. A reasonable model for low-density traffic is (25) with dq/dρ=c, where c is a constant.
    1. Show that ρ is constant along the (characteristic) curves x=ct+x0.
    2. If a car is alone on the highway, what does ρ(x,0) look like? What does ρ(x,t) look like?
    3. Explain why c represents the free speed of the highway.
  14. Solution
    1. Consider z(s)=ρ(x(s),y(s))=ρ(cs+x0,s). Then a direct computation gives

      z(s)=cρx(cs+x0,s)+ρt(cs+x0,s)=0.

      Thus, ρ is constant along with the line x=ct+x0.
    2. Suppose the center of mass of the car is located at x0 initially and the length of the car is L. Then, the density function is

      ρ(x,0)={1L,if |xx0|L2,0,if |xx0|>L2.

      According to (a), we know that

      ρ(x,t)={1L,if |xctx0|L2,0,if |xctx0|>L2.

    3. Based on the results of (a) and (b), we find that the car moves from x0 to x0+ct0 as t=0 becomes t=t0, which implies the velocity of the car is c.

  15. If ρmax denotes the maximum density of cars on a highway (i.e., under bumper-to-bumper conditions), then a reasonable relation between q and ρ is given by G(ρ)=cρ(1ρ/ρmax) where the constant c is the free speed of the highway (cf. Exercise 7). Suppose the initial density is

    ρ(x,0)={12ρmaxfor x<0ρmaxfor x>0.

    Find the shock curve and describe the weak solution. Interpret your result for the traffic flow.
  16. SolutionBased the setting in Problem, the partial differential equation is read as G(ρ)ρx+ρt=0. To solve the equation, we consider the characteristic curve x=x(s), t=t(s), z=ρ(x(s),t(s)) satisfying the following ordinary differential equation

    {dxds=G(z)=c(12zρmax),dtds=1,dzds=0 with the initial condition {x(0)=x0,t(0)=0,z(0)=ρ(x0,0).

    We can solve it by x(s)=cs(12ρ(x0,0)ρmax)+x0, t(s)=s, and z(s)=ρ(x0,0).
    • If x0<0, then the characteristic curve is vertical line x=x0.
    • If x0>0, then the characteristic curve is the line x=ct+x0 whose slope is 1c.
    In order to deal with the shock wave, we consider the ordinary differential equation

    ξ(t)=G(ρr)G(ρl)ρrρl=cρmax/4ρmax12ρmax=c2

    Thus, we get the shock curve x=ξ(t)=ct2 which passes (0,0). Hence, the weak solution is

    ρ(x,t)={12ρmax,if x<ct2,ρmax,if x>ct2.

    The figure of the solution can be shown as below.
    For the traffic flow, the region of the right side is crowded and the region of the left side is sparse. As time goes on, the crowded region extends to the left side.

  17. Using G(ρ) as in Exercise 8, describe the traffic flow after a long red light turns green at t=0; that is, the initial density is

    ρ(x,0)={ρmaxfor x<00for x>0.

    In particular, find the density at the green light, ρ(0,t), while the light remains green.
  18. SolutionAccording to Exercise 8, we can solve the ordinary differential equation with the initial condition (x(0),t(0),z(0))=(x0,0,ρ(x0,0)) by

    x(s)=cs(12ρ(x0,0)ρmax)+x0, t(s)=s, and z(s)=ρ(x0,0).

    • If x0>0, then the characteristic curve is the line x(s)=ct+x0 whose slope is 1c.
    • If x0<0, then the characteristic curve is the line x(s)=ct+x0 whose slope is 1c.
    Since the solution is undefined on the middel region, we need to consider the rarefaction wave. Let ρ(x,t)=v(x/t). Inserting it into the partial differential equaiton yields

    ct(12vρmax)v(xt)xt2v(xt)=0,

    which gives

    ρ(x,t)=v(xt)=(ctx)ρmax2ct=ρmax2(1xct).

    Hence, we obtain the weak solution

    ρ(x,t)={0,if x>ct,ρmax2(1xct),if ct<x<ct,ρmax,if x<ct.

    The figure of the solution is shown as below.
    When the red light turns green, the cars move from crowded to sparse, which makes the densitiy of the cars chagnes. In addition, we note that ρ(0,t)=ρmax/2.

沒有留言:

張貼留言