The Algorithm

In this algorithm, $s$ is a set of cells being processed, while $t$ is a set of cells to be processed in the next iteration. $l$ is the ``level'' of cells, or the distance from a destination. $d$ is the initial set of destination cells. $n(c)$ is a set of reachable neighbors of cell $c$. $v(c)$ is the value assigned to cell $c$, $v(c)$ is undefined for all cells initially.