Information technology for optimization of groundwater monitoring network using guided local search

M. S. Pavlov, S. V. Antonenko
2019 Actual problems of automation and information technology  
Розглянуто обчислювальну схему оптимізації мережі моніторингу стану підземних вод з використанням керованого локального пошуку на основі методу інтерполяції Сібсона. Ключові слова: керований локальний пошук, крігінг, інтерполяція Сібсона, метод природної околиці, оптимізація мережі моніторингу. Рассмотрена вычислительная схема оптимизации сети мониторинга состояния подземных вод с использованием управляемого локального поиска на основе метода интерполяции Сибсона. Ключевые слова: управляемый
more » ... альный поиск, интерполяция Сибсона, метод естественных окрестностей, оптимизация сети мониторинга. Optimization of the existing long-term groundwater monitoring networks can reduce the maintenance cost. This process should have minimal negative impact on the network in terms of monitoring information. Depending on the number of wells, finding the group of sites where measurements could be estimated with the lowest interpolation error can be a combinatorial problem. Combinatorial optimization problem can be solved by Guided Local Search algorithm. In this paper we consider the computational scheme for optimizing existing groundwater level monitoring network located in Kryvyi Rih, Dnipropetrovsk region, Ukraine, using a natural neighbor based guided local search. To find the groups of the redundant sites Python application for the network design optimization was developed. The solution was found for the network with 46 monitoring wells. The number of wells to remove was set to 7. The number of iterations was predefined. For guided local search number of iterations was set to 20. Number of iterations for local search was set to 100. Natural Neighbor interpolation requires the computation of the Voronoi diagram with interpolation point and without interpolation point, but the concept behind it is pretty straightforward. It is a weighted average interpolation technique, where the points used to estimate the value of an attribute at location ; are the natural neighbors of ;. By means the ones which share the edge of Voronoi cell with ;. Guided local search algorithm is trying to _________________
doi:10.15421/431911 fatcat:zpfa4ktkzra5bltaxzzwwt4sui