ocaml: phase 5.1 house_robber.ml baseline (max non-adjacent sum = 22)

Classic House Robber linear DP:

  dp[i] = max(dp[i-2] + houses[i], dp[i-1])

For [2; 7; 9; 3; 1; 5; 8; 6]:

  dp = [2, 7, 11, 11, 12, 16, 20, 22]
  max sum = 22

Optimal pick is {indices 0, 2, 5, 7}: 2 + 9 + 5 + 6 = 22 (all
non-adjacent).

Tests linear DP with early-return edge cases for n=0 and n=1,
inline-if rhs for max-of-two, dual look-back into dp[i-2] and
dp[i-1].

189 baseline programs total.

lib/ocaml/baseline/expected.json

@@ -78,6 +78,7 @@
   "hamming.ml": 4,
   "hanoi.ml": 1023,
   "hist.ml": 75,
+  "house_robber.ml": 22,
   "histogram_area.ml": 10,
   "huffman.ml": 224,
   "int_sqrt.ml": 1027,