AKL
et al.: MULTICELL CDMA NETWORK DESIGN
721
Fig. 12.
Network capacity for the uniform network and combined optimization
cases with and without the minimum capacity constraint.
Fig. 13.
Capacity in the 27-cell CDMA network, which is optimized using
base-station locations, pilot-signal powers, and power compensation factors
with a minimum capacity constraint.
shows, this is achieved at a small cost in total network capacity.
In fact, for relative user densities larger than four, the difference
between the total network capacity from combined optimization
and combined optimization (mc) is around three. We would like
to point out that, as Fig. 12 also shows, there is a larger drop
in total network capacity between the two cases of uniform net-
work and uniform network (mc). Adding the minimum capacity
constraint in (35) without the optimization of PCFs, pilot signal
powers, and base-station locations causes a larger reduction in
the total network capacity. We should also point out that the
amount of reduction depends on the number of hot spots in the
network. An increase in the number of hot spots and an increase
in the relative user density in the hot spots causes a greater re-
duction in the total network capacity.
Finally, the capacity of the cells for the relative user density
of five is given in Fig. 13 in parentheses. The smallest capacity
in any cell is 17, as opposed to 13 in Fig. 7. The MIP solution
yields a network capacity of 564, which is only one less than
the capacity achieved previously using (34). Thus, our results
show that maximizing the network capacity, with a minimum
capacity constraint, by varying the PCFs, base-station locations,
and pilot-signal powers, is the best way to increase capacity in
the cells individually and in the network as a whole.
VII. C
ONCLUSION
We show how to increase the reverse-link capacity in a
CDMA network by varying the transmission power of the
mobiles, the pilot-signal powers, and the base-station locations.
We calculate the derivative of the reverse-link network capacity
with respect to pilot-signal powers, base-station locations, and
power compensation factors. These derivatives are then used in
an optimization procedure to maximize the network capacity.
The results confirm that for a uniform user distribution, a
uniform network layout with equal-sized cells is optimal.
For a nonuniform distribution, more cells need to be located
inside the hot-spot cluster. If pilot-signal power is the only
variable parameter, then an increase in pilot-signal powers of
congested cells increases network capacity. Even though the
intracell interference increases, a greater reduction in intercell
interference is achieved, which yields an increase in the overall
capacity. We also construct and solve constrained optimization
problems, which guarantee a minimum capacity for every indi-
vidual cell while maximizing the total network capacity. These
results indicate that including a hard constraint on the minimum
capacity of individual cells has little effect on network capacity
given the flexibility of optimizing the transmission power
of the mobiles, the pilot-signal powers, and the location of
the base stations. However, without such flexibility, the hard
constraint on cell capacity imposes a significant penalty on
network capacity. The network design technique introduced
accommodates postdeployment design changes in response to
changes in demand, particularly by changing the PCFs and the
pilot-signal powers.
R
EFERENCES
[1] K. S. Gilhousen, I. M. Jacobs, R. Padovani, A. J. Viterbi, L. A. Weaver,
and C. E. Wheatley, “On the capacity of a cellular CDMA system,”
IEEE