The authors presents a study of the effect of correlation among the multi-hop path gains on the average channel capacity and the outage performance of a power line communication system equipped with decode-and-forward relays. The path gain of each hop follows a log-normal distribution and the joint density is modelled by a multivariate log-normal distribution with exponential correlation. A distance dependent signal attenuation factor is added to the channel model to incorporate the effect of physical distance between the nodes. Furthermore, a Gaussian mixture (Bernoulli-Gaussian) noise is included in the system model to analyse the effect of impulsive noise together with thermal noise. Approximate closed-form expressions for the end-to-end average channel capacity and the outage probability are derived. Numerical results showing the impact of the amount of correlation, the number of hops, the impulsive noise power, and the spread of the average power of the path gains on the performance are presented. The results show that for a fixed transmit power and end-to-end distance between source and destination, the performance improves with increasing number of hops; however, the amount of improvement reduces with increasing channel correlation.