Representative Sample Selection via Random Search with Application to Surveying Communication Lines

The Municipal Transport Union of the Upper-Silesian Industrial District (KZK-GOP) organizes the public transport system in 23 municipalities of the Silesian Region, covering the area of 1400 square kilometers with over two million inhabitants. Its main sources of financing include the fare income as well as subsidies from member municipalities. Since the year 2008 the amount of subsidy from each member depends on the profitability of communication lines that cross its territory. To calculate the profitability passenger loads on these lines must be assessed by means of a sample survey. Most lines in the system cross the area of several municipalities. Moreover, specific requirements concerning the composition of the line sample must be met to ensure the representativeness of estimates and preserve the credibility of the survey. This makes the use of simple random sampling designs problematic. In this paper a specialized random search algorithm for selecting representative samples is presented. It incorporates random sampling as well as purposive selection of sample units in a, way that makes the evaluation of sampling errors possible. The procedure attempts to minimize them.