31 lines
676 B
Markdown
31 lines
676 B
Markdown
|
# Search modes
|
||
|
|
|
||
|
|
## Heuristic
|
||
|
|
|
||
|
|
A*, Greedy and Beam search all use the following heuristic to select candidates to expand on the next layer of the graph
|
||
|
|
|
||
|
|
$$mult(n) =
|
||
|
|
\begin{cases}
|
||
|
|
4 &\text{if $n$ is a neutron star} \\\\
|
||
|
|
1.5 &\text{if $n$ is a white dwarf star} \\\\
|
||
|
|
1 &\text{otherwise}
|
||
|
|
\end{cases}
|
||
|
|
$$
|
||
|
|
|
||
|
|
$$d(a,b) = \sqrt{(a_x-b_x)^2+(a_y-b_y)^2+(a_z-b_z)^2}$$
|
||
|
|
|
||
|
|
$$
|
||
|
|
h(\text{node},\text{goal})=
|
||
|
|
\max(
|
||
|
|
d(\text{node},\text{goal})-
|
||
|
|
(
|
||
|
|
\text{range}*mult(\text{node})
|
||
|
|
),0)
|
||
|
|
$$
|
||
|
|
|
||
|
|
potential new heuristic:
|
||
|
|
|
||
|
|
$$
|
||
|
|
h(\text{node},\text{next},\text{goal})=
|
||
|
|
1 - {\cos^{-1}(|(\text{next}-\text{node})| \cdot |(\text{goal}-\text{node})|)\over\pi}
|
||
|
|
$$
|