as next generation wireless and mobile networks continue to evolve, the scarcity of frequency spectrum and bandwidth are on the rise. while ultra-dense small-scale networks aim to satisfy the demand for ultra-high data rate needs, they offer limited coverage to mobile users. while relay-based network coverage expansion emerges as an appealing solution, it typically contributes to the additional transmission delay between a source node (sn) and a destination node (dn). to support high data rate communication and reduce the transmission delay, the heterogeneous frequency bands could be adaptively/simultaneously used to transmit the data packets over the relay node/s (rn/s). in this thesis, the exploitation of heterogeneous frequency bands in this manner is referred to as the multi-band communication. in this thesis, the advantages and shortcomings of relay-based transmission with multiband communication are studied in a systematic manner. by first formulating the theoretical problem, it is then investigated that how to develop efficient approaches and algorithms to effectively reduce the packet relay latency. in particular, the primary concern is the added delay due to relaying data packets from sn till dn via one or more rns and any additional coordination among nodes might end up in additional computational overhead. therefore, there emerges a trade-off between optimal (i.e., best) and fast decisions. to address the trade-off between the quality and speed of the required solution, both centralized and distributed methods are analyzed and proposed, so as to reduce the packet latency from sn to dn. in centralized optimization approach, as the number of rns increase, an optimal solution is difficult to determine because of non-deterministic computation time. [...]