14.5 Path Integrals, Revisited

We've defined path integrals for rulers that are differentials. Unfortunately, this definition is useless for many rulers: a wild ruler can never be written as the differential of a function \( z \depends (x, y) \). To work with wild rulers, we'll need to broaden our definition of the path integral.
Definition.
Suppose we have a ruler, a path, and a time interval as follows.
\[ \begin{align} r &: \Ruler_{(x, y)} & p &: \Path_{(x, y)} & i &: \Interval_{t} \end{align} \]
We calculate the path integral
\[ \begin{align} \Bigl\langle{r}\mathrel{\overset{p}{\textstyle{\int}}}{i}\Bigr\rangle &: \Number \end{align} \]
by using the path \( p \) to rewrite the path integral as a time integral.
By moving path integrals from \( (x, y) \)-coordinates to time \( t \) we can take many more integrals!
Example.
Let's compute the path integral
\[ \begin{align} &\Bigl\langle{r}\mathrel{\overset{p}{\textstyle{\int}}}{i}\Bigr\rangle \end{align} \]
with ruler, path, and interval given as follows.
\[ \begin{align} r &: \Ruler_{(x, y)} & p &: \Path_{(x, y)} & i &: \Interval_{t} \\ r &= y \mult \bar{\standard}_{x} + (2 y - x) \mult \bar{\standard}_{y} & &\left\{\begin{aligned} x_p &= t^2 \\ y_p &= t \end{aligned}\right. & i &= [0, 3]_{t} \end{align} \]
You can check that the ruler \( r \) is wild. So there is no hope of finding a function \( z \depends (x, y) \) that satisfies
\[ \begin{align} \bar{\diff} z &= y \mult \bar{\standard}_{x} + (2 y - x) \mult \bar{\standard}_{y}. \end{align} \]
Let's instead use our path \( p \) to rewrite \( r \) as a ruler that measures time.
\[ \begin{align} r &= y \mult \bar{\diff} x + (2 y - x) \mult \bar{\diff} y \\ &= t \mult \bar{\diff} \bigl(t^2\bigr) + \bigl(2 t - t^2\bigr) \mult \bar{\diff} t \\ &= t \mult 2 t \mult \bar{\diff} t + 2 t \mult \bar{\diff} t - t^2 \mult \bar{\diff} t \\ &= \bigl(2 t^2 + 2 t - t^2\bigr) \mult \bar{\standard}_{t} \\ &= \bigl(t^2 + 2 t\bigr) \mult \bar{\standard}_{t} \end{align} \]
To solve our integral, we should look for a function \( z \depends t \) that satisfies
\[ \begin{align} \bar{\diff} z &= \bigl(t^2 + 2 t\bigr) \mult \bar{\standard}_{t}. \end{align} \]
By anti-differentiating, we find
\[ \begin{align} \bigl(t^2 + 2 t\bigr) \mult \bar{\standard}_{t} &= \bigl(t^2 + 2 t\bigr) \mult \bar{\diff} t \\ &= t^2 \mult \bar{\diff} t + 2 t \mult \bar{\diff} t \\ &= \frac{1}{3} \mult 3 t^2 \mult \bar{\diff} t + 2 t \mult \bar{\diff} t \\ &= \frac{1}{3} \mult \bar{\diff} \bigl(t^3\bigr) + \bar{\diff} \bigl(t^2\bigr) \\ &= \bar{\diff} \biggl(\frac{1}{3} \mult t^3\biggr) + \bar{\diff} \bigl(t^2\bigr) \\ &= \bar{\diff} \biggl(\frac{1}{3} \mult t^3 + t^2\biggr). \end{align} \]
We can calculate our path integral as a time integral.
\[ \begin{align} \Bigl\langle{r}\mathrel{\overset{p}{\textstyle{\int}}}{i}\Bigr\rangle &= \biggl\langle{y \mult \bar{\standard}_{x} + (2 y - x) \mult \bar{\standard}_{y}}\mathrel{\overset{p}{\displaystyle{\int}}}{[0, 3]_{t}}\biggr\rangle \\ &= \biggl\langle{\bigl(t^2 + 2 t\bigr) \mult \bar{\standard}_{t}}\mathrel{\displaystyle{\int}}{[0, 3]_{t}}\biggr\rangle \\ &= \biggl\langle{\bar{\diff} \biggl(\frac{t^3}{3} + t^2\biggr)}\mathrel{\displaystyle{\int}}{[0, 3]_{t}}\biggr\rangle \\ &= \biggl\langle{\frac{t^3}{3} + t^2}\mathrel{\displaystyle{\int}}{\boundary [0, 3]_{t}}\biggr\rangle \\ &= \biggl\langle{\frac{t^3}{3} + t^2}\mathrel{\displaystyle{\int}}{[3]_{t}}\biggr\rangle - \biggl\langle{\frac{t^3}{3} + t^2}\mathrel{\displaystyle{\int}}{[0]_{t}}\biggr\rangle \\ &= \biggl(\frac{27}{3} + 9\biggr) - (0 + 0) \\ &= 18 \end{align} \]
We've found our path integral!