This paper describes a novel negotiation protocol for cellular networks, which intelligently improves the performance of the network. Our proposed reactive mechanism enables the dynamic adaptation of the base stations to continuous changes in service demands, thereby improving the overall network performance. This mechanism is important when a frequent global optimization is infeasible or substantially costly. The proposed local negotiation mechanism is incorporated into a simulated network based on cutting-edge industry technologies. Experimental results suggest a rapid adjustment to changes in bandwidth demand and overall improvement in the number of served users over time. Although we tested our algorithm based on the service level, which is measured as the number of covered handsets, our algorithm supports negotiation for any set of parameters, aiming to optimize network's performance according to any measure of performance specified by the service provider.