Option B.
Let’s say there’s a production rule where all symbols on R.H.S are terminals. So to convert that into CNF we’ll have maximum $\left | T \right |$ new variables and hence maximum $\left | T \right |$ new production rules.
After that we’ll be left with all variables on the R.H.S of a production rules. If we have A → BCD we can do like,
A→ VD , V → BC.
If we have A→ BCDE , we can convert that into A→ VE , V→ UD , U→ BC. Notice that for ‘n’ number of variables we can have a maximum of (n – 1) production rules.
Hence for ‘K’ symbols on the R.H.S we’ll have (K – 1) production rules for each production rule in G.
so for $\left | P \right |$ productions we can have a maximum of (K – 1) $\left | P \right |$ rules. This added with $\left | T \right |$ will make the final answer as
(K – 1) $\left | P \right |$ + $\left | T \right |$