References
See: Hägerstrand, Torsten, “Statistiska primäruppgifter, flygkartering och ‘Data Processing’ — maskiner; Ett kombinationsprojekt,”Svensk Geografisk Årsbok, 1955; and Nordbeck, Stig, “Location of Areal Data for Computer Processing,”Lund Studies in Geography, Series C 2, Lund, 1962, and “Framstälning av kartor med hjälp av siffermaskiner,” Lund, 1964.
Provided, of course, that there is no direction within the town which is more probable than any other. Thus the town must not have a marked linear form or in any other way be irregular in shape, It should perhaps be emphasized that this extension constant can not be used in a given particular case, where the real distance between two points has to be calculated, but only where average values have to be calculated, or in similar instances where there is good reason to suppose that possible errors will cancel each other out.
If only ten distances are included,q=1.20, a result which does not differ much.
Figure 4 shows certain distances which giveq ν of about 1.40, although these are rather long. This is because of a small barrier within the district: a stream with only a few bridges. These distances influence but little the final value ofq, but if only the first ten distances are included in the calculations,q=1.26. This suggests that small barriers need not be given much consideration. If, on the other hand, the barriers are major, and are a source of greater hinderance, ignoring them would result in a very distortedq value.
Of course, it is possible to add conditions: e. g., in the investigation of routes between two pointsP a andP b , only the barrier nearest to pointP a need be taken into consideration. Further, to make the programs less intricate, it is necessary, to delineate the barriers schematically in the from of straight line segments or as simple polygons.
If a node pointK ν has more than four neighbours, it has to be a neighbour of itself and to have the next number ν+1 too. The distance between these “two” pointsK ν andK ν+1 is 0, of course.
Since, in cases where no one-way roads exist, the distance from one node point ν to another node point number μ is equal with the distance from point μ to point ν, the storage matrix needs only (n−1)2/2 cells if there aren node points in the road net.
In data schedules with nodes having three or more neighbours, the pointsN 1 andN 2 have to be written as immediate neighbours.
The programs discussed in this paper are known as: 1. NORAS. The program investigates whether the the straight line between two points is barred or not; this is thus a sub-routine of the other barrier programs. 2. NORASP. The program which analyses routes where the barrier is a polygon. 3. NORAN. The shortest route program.
Author information
Authors and Affiliations
Rights and permissions
About this article
Cite this article
Nordbeck, S. Computing distances in road nets. Papers of the Regional Science Association 12, 207–220 (1964). https://doi.org/10.1007/BF01941253
Issue Date:
DOI: https://doi.org/10.1007/BF01941253