Hi Aritra,
The major problem with the approach that I can see is that it degrades to proof-of-work -- attackers with computational resources can produce blocks by brute-forcing the proof-of-work needed to create addresses and thus sybil the network by self-generating enough transactions to trigger block production. A second problem is that no-one is incentivized to propagate transactions.
If you are thinking along these lines, you might find Saito (http://saito.tech) interesting. It uses a proof-of-transactions approach that measures the work done by block producers in terms of the transaction fees they have collected. It is similar in the sense that it eliminates mining, but avoids these attack vectors.