Suppose that we are looking for a function that solves the PDE
The graph of a solution to (7.10) is a surface in the -space defined by the equation . The gradient vector is normal to this surface at each point. Noting that the PDE (7.10) can be equivalently written as
we conclude that the vector is everywhere orthogonal to the normal , hence it lies in the tangent space to the solution surface. This is the geometric interpretation of the quasi-linear PDE (7.10): the vector field is tangent to the solution surface at each point. The system of ODEs associated with this vector field can be written as
The characteristic curves ``fill" our solution surface ; skipping ahead to Figure 7.2 may help the reader visualize this situation. As an example, it is useful to think about the case when . Then (7.10) says that the derivative of in the direction of the vector is 0, implying that stays constant along solutions of the equations , . Every function whose level sets are solution curves of these two equations in the -plane is a solution of the PDE, and the characteristics are horizontal slices of the corresponding solution surface. Of course, in general (for nonzero ) characteristics are not horizontal.
Now, in principle we should be able to solve the characteristic ODEs and obtain the characteristics. How does this help us solve the original PDE (7.10)? Without additional specifications, there are ``too many" characteristics to identify a desired solution surface ; indeed, there is a characteristic passing through every point in the -space, so they fill the entire space. We know that for ODEs, to fix a particular solution we must specify a point through which it passes (an initial condition). The counterpart of this for the PDE (7.10) is that we must specify a suitable curve in the -space through which the solution surface passes; we then obtain what is called a Cauchy (initial value) problem. Geometrically, the idea is that characteristics that pass through points on this initial curve will determine the solution surface of the PDE, as illustrated in Figure 7.2 (with the thick curve representing an initial curve).
An initial curve can be defined in parametric form as
where . Then, the desired characteristics--i.e., the solutions of the characteristic equations (7.11) whose initial conditions lie on this initial curve--are given by
is nonsingular along the initial curve. What is the geometric significance of this latter condition? The columns of the above Jacobian matrix are the -components of the tangent vectors to the initial curve and to a characteristic, respectively, and they must be linearly independent. In other words, when projected onto the -plane, the characteristics and the initial curve must be transversal to each other. We conclude that under this transversality condition, the Cauchy problem for (7.10) is (locally) uniquely solvable, at least in principle, via the method of characteristics. To calculate the actual value of for some specific and , we need to find a characteristic that connects the initial curve to a point with these -coordinates and then integrate the last equation in (7.11) along that characteristic segment. We will discuss this procedure in more detail later for the case of the HJB equation.
Characteristics can also be defined for the general PDE (7.9) which for takes the form
Then the characteristic strips are defined by the following equations:
In the above discussion we assumed that the independent variable lives in the plane (i.e., ). However, it is straightforward to write down the equations of the characteristic strips for a general dimension : we must simply replace the indices in (7.14) by . In the next subsection we will work in this more general setting.