The article describes a method for solving the problem of optimal object allocation using evolutionary algorithms. An approach using network theory as a tool for describing the allocation problem is presented, providing a way to represent the problem with or without the Markov property. At the same time, the problem of single- or multi-criteria network optimization is defined and described using examples. Subsequently, optimization calculations were performed for two examples of single- and dual-criteria object allocation described by networks without and with the Markov property, respectively. The results were compared with the systematic search method. The obtained results provide a basis for stating the high utility of evolutionary algorithms in solving allocation problems.One of the more rational approaches to the object allocation problem is the use of network theory, in which possible solutions are represented by network arcs, while optimal allocation boils down to finding the best path in the network.
Networks representing the allocation problem can have different characteristics, i.e., there may be networks possessing the so-called Markov property or lacking this property, and allocation can take place based on one criterion or multiple criteria, resulting in either a single solution or a set of Pareto optimal solutions [3]. (…)
The article contains 19179 characters.
Source: Logistyka Magazine
Optimal object allocation using evolutionary algorithms – PDF file for download







