Cryptosystems constructed using lattice are worst-case hardness.
Actually, ECC also has the 'worse-case hardness' property; it is easy to blind an ECC DLog problem, converting an arbitrary case into a random case (and hence the average hardness of DLog problems is also the worse-case hardness).
Can we mention, cryptosystems based on lattices are simple to implement as it simpler operation(matrix multiplications) compared to ECC?
Well, to be efficient, lattice-based systems often use fancier matrix multiplication algorithms than the simple textbook algorithm; this cuts into the simplicity somewhat.
One advantage that you don't list for lattice-based systems: we know how to build somewhat homomorphic and fully homomorphic systems based on Lattice techniques; we don't know how to do that based on ECC.
On the other hand, I'm not exactly what you mean by "How can we justify"; diversity of hard problems isnt' a bad thing. It may be that we figure out how to efficiently compute discrete logs; actually, we're pretty certain that large scale Quantum Computers will be created at some point (even if we don't know if that's 10 years or 30 years down the road). To address that eventuality, looking at cryptosystems based on other hard problems (and believed to be quantum safe) is certainly justified.