1: \begin{abstract}

2: %\boldmath

3: This paper discusses a special type of multi-user communication

4: scenario, in which users' utilities are linearly impacted by their

5: competitors' actions. First, we explicitly characterize the Nash

6: equilibrium and Pareto boundary of the achievable utility region.

7: Second, the price of anarchy incurred by the non-collaborative Nash

8: strategy is quantified. Third, to improve the performance in the

9: non-cooperative scenarios, we investigate the properties of an

10: alternative solution concept named conjectural equilibrium, in which

11: individual users compensate for their lack of information by forming

12: internal beliefs about their competitors. The global convergence of

13: the best response and Jacobi update dynamics that achieve various

14: conjectural equilibria are analyzed. It is shown that the Pareto

15: boundaries of the investigated linearly coupled games can be

16: sustained as stable conjectural equilibria if the belief functions

17: are properly initialized. The investigated models apply to a variety

18: of realistic applications encountered in the multiple access design,

19: including wireless random access and flow control.

20: \end{abstract}

21: