Traffic flow quantities can be conveniently represented by an array. The place of Origin is represented by the array row. The palce of Desinations by the array column. The complete array is an OD table. It may be for a specific period of time, purpose, or other classification. Ideally it will show the traffic from any origin to any destination.
Each table is a n x m array, where n is the number of origin and m the number of destination zones of the region. If the origin and destination zones are the same then n=m and the usual square OD table appears.
A time, series or other patterns therfore requires a table for each disag- gregation such as for: morning flows passengers, commodities, purposes, etc.
To use the OD table for disaggregate and/or time series models requires that the zone structure remain fixed for the model life. The array will have at least three dimensions: Origin, Destination, other disaggregation dimension. There may be more than one disaggregation dimension such as combinations of: commodity, mode, purpose, time period.
A multidimensional array can be manipulated by software including APL.
The major difficulties with OD tables is obtaining accurate data if actual measured information is required. Various model can be used to estimate an OD table but they will only be estmates. The major mdelling techniques are: Gravity Models, Optimization Models, and Growthing of Historical data. The general term is Trip Distribution.
End to date, ams 990824