<FONT face=Verdana>In this paper

we study the power allocation problem in decode-and-forward coop erative diversity systems.We first present the mathematical model of the power allocation problem

which can minimize the total power consumption of the source and partner under the average bit error rate (BER) constraints at the partner an d the destination.Accordingly

the model not only can exploit the inherent spati al diversity in multiuser wireless systems

but also can reduce both the energy c onsumption of mobile nodes and the cochannel interference in wireless networks. We then derive the BER expressions at the partner and the destination which can characterize the performance of cooperative diversity systems over multipath fad ing channels.According to the expressions

we propose a distributed power alloca tion algorithm by which the source

the partner and the destination can compute t he optimal power allocation scheme (OPAS) on a message-by-message basis.The p roposed algorithm not only has little overhead but also adapts to the changes of channel state.Therefore

the algorithm can conform to the decentralized and dyna mic nature of wireless networks.In addition

we present simulations to confirm th e proposed algorithm.