The simulation model adaptability to real network behavior is the key concept in wireless communications. In a cellular network, many procedures such as call admission, hand-off, etc take place simultaneously for every individual user. Every network procedure acts autonomously, interacts with the network environment (gathers information such as interference conditions), takes decisions (e.g. call establishment), etc. Although this is known in the literature, there is lack of suitable representations for such network procedures in the simulation systems proposed so far, thus compromising simulation model adaptability to real network behavior. To achieve such adaptability we herein propose to change the point of view in network procedure representation. Instead of viewing them as independent programming functions or even objects in a high level language, which are sequentially executed, due to their aforementioned properties it is proposed that such network procedures could be more efficiently modeled as agents. Considering this new approach, the agent cooperation and communication in terms of negotiation and agreement is a critical issue. In this paper we present a centralized cooperative multi-agent negotiation scheme applied to a multi-agent layered architecture for designing and simulating resource allocation in cellular communication systems, based on organizational modeling. Moreover, we show the way that the rules and implementation methods of agent negotiation affect the adaptation grade of simulation model to the real cellular network behavior.