Coupling developmental rules and evolution to aid in planning network growth

In order to meet the increasing demand for capacity on BT's telecommunicati ons networks, they must be grown. Planning suck growth in order to cost eff ectively meet the projected demand is thus an important concern. This paper describes a tool that was developed to aid in such a task. The system empl oys a simulation of both the network and the demand that must be satisfied. A set of planning rules are defined which are triggered bq, conditions in the simulated network and result in the addition or reconfiguration of hard ware in an attempt to better accommodate the simulated demand A genetic alg orithm is employed to evolve the parameters of these rules with the overall effect of optimising the resulting networks against constraints of cost an d quality of service. In this way, much like the real world, evolution is a modifier of the dynamics of a developmental process. Application of the to ol is demonstrated on a specific instance of network growth - the dial-up I P access network that requires rapid expansion.