One of the question in my resent code challenge was related to a physical algorithm. I’ve been explained ahead of time that during the test I will be asked to write an algorithm to solve a problem related to the physical world. e.g. given the topology of a mountain pass as an array, write an algorithm to figure out the cheapest way to build a road over the pass.

First what came to my mind after reading that was Dijkstra’s algorithm also known as Dijkstra’s Shortest Path First algorithm or, SPF algorithm. In many real world situations we are looking for the efficient solution - optimal distance or cheapest cost. Conceived by computer scientist Edsger W. Dijkstra in 1956 and published three years later the algorithm that is used to calculate the shortest path between two nodes on a graph, now days even has applications in routing network traffic, machine learning, and artificial intelligence.

For the algorithm to work, the graph must meet three conditions:

- Edges have non-negative weights
- Weights are directional A->B may be not equal B->A
- The graph is acyclic, after moving A->B, we will not consider moving back to vertex B from vertex A as this would create an infinite number of routes

The algorithm would take the Graph and starting and eding point.

```
function Dijkstra(Graph, start, end):
2
3 create vertex set Q
4
5 for each vertex v in Graph:
6 dist[v] ← INFINITY
7 prev[v] ← UNDEFINED
8 add v to Q
10 dist[start] ← 0
11
12 while Q is not empty:
13 u ← vertex in Q with min dist[source]
14
15 remove u from Q
16
17 for each neighbor v of u:
18 alt ← dist[u] + length(u, v)
19 if alt < dist[v]:
20 dist[v] ← alt
21 prev[v] ← u
22
23 return dist[], prev[]
```

A more general problem would be to find all the shortest paths between source and target (there might be several different ones of the same length). Then instead of storing only a single node in each entry of prev[] we would store all nodes satisfying the relaxation condition.

One of the variations of this algorithm uses a priority queue. Such data structure leads to a faster computing time. Algorithm in Javascript could be found here.

It turned out that the problem I was given was much simplier that I had been preparing for and the Dijkstra’s Algorithm for that task would be an overkill that would’ve made a solution more complex than it should be.

Here is the problem:

Here is my solution to the problem:

```
function solution(A) {
let castleCount=2
let start=0
let current=A[0]
let i=0
while( i<A.length){
while(current==A[i]) i++
if (start!==0 && i!==A.length-1){
if (A[start-1]>A[start]&&A[i]>A[i-1] || A[start-1]<A[start]&&A[i]<A[i-1] ){
castleCount++
}
}
if (start===0 && i===A.length){
castleCount--
}
start=i
current=A[i]
i++
}
return castleCount
}
```

The code challenge was timed - two hours sharp - and the problems were not given in advance which means that I had to understand the problem, come up with solution ideas and implement them witin the given time. That was one of the three problems given at the code challenge with an expectation to finish them all within two hours given. Two other problems were to gebug a ticky code with JS closure and to implement a math algorithm with recursion involved. And that was only the pre-screening stage in the interview process. Tells a lot about current industry standarts.