In this paper, a system of GIG/l/K queue is considered. The optimal
system's capacity (K), when the system is optimized with respect to the benefit
of the entire system (social optimization) and when the criterion for optimality
is individual gains (individual optimization), is determined and compared. In
social optimization, the system capacity is obtained through maximization of
the system's profit. However, when individual gains is the criteria, the system
capacity is determined as a result of customers not joining the system because
they estimate that getting service from the system does not yield them any
profit and therefore leave the system. It is shown by simulation that irrespective
of the traffic intensity, p and arrival and service time distributions with
different failure and acceleration rates, the K obtained from social optimization
of the system is always equal or less than the K obtained from individual
optimization. Thus, the social optimization of the system not only maximizes
the system's profit but also preserves individual gains and prevents the
customer from leaving the system. In this paper, some other interesting results
are also reported.