In order to analyze the concept of network (node) efficiency we offer a mathematical model for identifying choke points.
The underlying theory is based on the relationship between link flow and time interval. We will focus on analyzing network
conditions beyond the equilibrium state (i.e. when the network begins to disintegrate and becomes less reliable). If flow a link flow qa, over the time interval t0 (this can be selected based on industry practice for data collection) is to pass through a node N, the arrival rate ba over that time interval is given by,
Where, = Initial link free flow per unit time = incidental change in link flow due to additional system loading within
interval.
If distribution function (given by ratio of link flow to total outflow) is known, we can predict the loading characteristics of
the succeeding links by computing the departure rates from the node. Equation (1) can be re-written as,
The term will change due to the state of the network: free flow versus jammed condition.
Under free flow conditions, the outflow split to succeeding arcs is somewhat stable as user choices are unconstrained.
However, as disaster and evacuation activities build, some links become undesirable to the user.
This coincides with increasing values of beyond the stability state and as departure rate, from node approaches zero.
We call this the desirability index of the link and as its values approach 1, the more the likelihood of user to choose an alternate path.
Desirability index is an indicator of potential choke points. Thus, if flow conservation is maintained through any given node N, Equation (5) below can be written to show that inflow to node ( ) must equal to outflow from node.