Different metrics are used for measuring the goodness of packet routing between the source and destinationpairs in a communication network. One approach is to compute the paths from the measured metric values, which arepredetermined independently according to the network resources. The combination of some metrics of different naturesto get a composite cost function for the routing process is non-trivial. We consider two metrics, namely delay andpacket loss rate. The first one takes values from the large numbers and the second from the small. We propose somemethods for constructing composite functions of these metrics without any constraints and use them for the shortestpath calculations. We give the numerical results of the proposed composite functions versus the Dijkstra’s algorithmwith the individual metrics. We spot out the best function according to our computer simulation results. Our compositefunction works for any arbitrary point-to-point networks. To the best of our knowledge, the technique is novel. Ourresults show an efficient way to balance the effects of the metrics in the context of many-to-many routing.