RPL is a standard routing framework for low-power and lossy networks (LLNs). LLNs usually operate in challenged conditions, therefore RPL can be adapted to satisfy requirements of a particular LLN. RPL facilitates this through objective functions (OFs). An OF is used to discover and maintain data forwarding paths based on the requirements of LLNs. In RPL, different OFs can use different routing metrics in different ways. In this paper, we design different OFs and analyse their impact on RPL performance in multi-gateway ad-hoc LLNs. In conjunction with the shortest hop-count, our designed OFs also use the following tie-breaking metrics: available bandwidth, delay, buffer occupancy, and ETX. Our OFs use the tie-breaking metrics on a greedy or an end-to-end basis. In our experimental analysis, we consider the impact of duty-cycling, number of gateways, and data traffic load on the OFs’ performance. Our results demonstrate that, generally speaking, the performance improves with an increase in the number of gateways. In the absence of duty-cycling, the greedy approach is better compared to the end-to-end approach, and using delay, buffer occupancy, and ETX metrics as the tie-breaking metrics in conjunction with the shortest hop-count metric yield the best performance. In a relatively high data traffic load, all OFs perform similarly. In duty-cycling mode, frequent changes in the parent node incur extra synchronization time between a sender and receiver. OFs that use the tie-breaking metrics on an end-to-end basis do not frequently change parent nodes, hence they demonstrate better performance. Furthermore, in duty-cycling mode, the shortest hop-count metric demonstrates the best performance compared to the other metrics.