Multipath transmission of raw sensor signals is the customary technique used in the wireless sensor network to improve end-to-end delivery. However, this technique suffers significantly because of the occurrence of multiple copies of data at the destination and their collision. The Collision-Free Nearest Neighbour Assertion (CNNA) method with n-d tree structure improves the collision removal which, in turn, avoids duplicate packets, but load balancing among neighbouring nodes is an essential issue. Optimising network performance by considering various network parameters and load balancing the network demands a good evolutionary-based optimisation technique other than traditional algorithms. Optimisation techniques based on Particle Swarm Optimisation (PSO) and Genetic Algorithm (GA) are applied and compared against various network parameters in this work.