Decentralized optimization of the flexible production lines

Abstract

The shortening of the production cycle and increasing impact of the technological innovations evokes improvement of the methods used in the production line scheduling. The aim of the presented research is a proposal of the decision model that enables a flexible reaction to the changing production conditions. The central decisions are substituted by the decisions performed on the independent machines level. The machines utilize rather restricted information on the capacities utilization of their technological neighbours. The decisions follow the decision tables (decision chromosomes) re‐coded in the course of the evolution. The Genetic algorithms standing behind the model, enable identification of the highly acceptable solutions that is proved by a set of the simulation experiments. The set of independent machines becomes self‐organized, having a significant positive effect on the production line capacities utilization. The decentralization aspect makes our proposal somewhat different from other research in the field. The philosophy is built on the fact that each machine makes its production decision based on the available information, which it has at its disposal in a given time. The machine flexible reacts to its neighbours' capacity utilization (machine before and after given production machine) in the production process flow.