Intelligent Support for Resource Distribution in Logistic Networks Using Continuous-Domain Genetic Algorithms

The paper addresses the issue of improving the goods distribution efficiency in logistic networks subjected to uncertain demand. The class of networks under consideration encompasses two types of entities - controlled nodes and external sources - forming a mesh interconnection structure. In order to find the optimal operating conditions for the a priori unknown, time-varying demand, numerous, computationally involving simulations need to be conducted. In this work, the application of genetic algorithms (GAs) with continuous domain search is proposed to optimize the goods reflow in the network. The objective is to reduce the holding costs while ensuring high customer satisfaction. Using a network state-space model with a centralized inventory management policy, GA automatically adjusts the policy parameters to a given network topology. Extensive tests for different statistical distributions validate the analytical content.