Adaptation of the probability changing method for Weber problem with an arbitrary metric (Q6486674)

The author considers the planar multi-facility Weber problem with restricted zones and non-Euclidean distances. He proposes an algorithm based on the probability changing method and proves its efficiency for approximately solving the Weber problem by replacing the continuous coordinate values with discrete ones. An example for a problem with a discrete coordinate system \(200\times 400\) is given. A version of the described algorithm for multiprocessor systems is also presented. The proposed probability changing method can be used for the approximate solution of the planar generalized Weber problem with an arbitrary monotone non-Euclidean metric. The computational results indicate that modern computer facilities (multiprocessor systems, inexpensive clusters) allow for solving such problems with the appropriate accuracy.