The presented paper is dealing with knowledge discovery from data of the stochastic production plan to determine an adequate inventory control algorithm on a selected time interval with regard to given cost function and restrictions. In this preliminary research, machine learning methods, e.g. neural networks, decision trees and Bayes classifier, were used. Methods were learning on production plan characteristics, e.g. mean and variance, and periods, determined by Fourier analysis. The classification accuracy is presented together with comparison of machine learning methods.
B.03 Paper at an international scientific conference
COBISS.SI-ID: 6066451Web based decision support information system was developed to support the purchasing logistics with a goal of optimizing inventory control and reducing costs. User interface was implemented on Apache web server with web programming languages PHP and Java Script. System is supported by MS SQL Server 2008 database. Inventory control computational core was implemented with C# programming language and is based on simulation model and inventory control algorithms.
F.15 Development of a new information system/databases
COBISS.SI-ID: 6626323The user manual was made with a goal of training the staff in purchasing logistics to use the decision support simulation system in their every day routine. The manual incorporates detailed description of all system functionalities. It is supported with graphical material and detailed description of particular procedures for easier understanding.
F.17 Transfer of existing technologies, know-how, methods and procedures into practice
COBISS.SI-ID: 6626579The project leader has been awarded with the Trimo research award for the doctoral dissertation that is a foundation for this postdoctoral research project.
E.01 National awards
COBISS.SI-ID: 5472019The simulation models are used for decision support and learning in enterprises and in schools. Cases of successful applications demonstrate usefulness of weak anticipative information. Job shop scheduling production with a makespan criterion presents a real case customized flexible furniture production optimization with genetic algoritms Simulation based system for products with stochastic lead times and demand describes inventory optimization with a goal of reducing costs without producing stock-outs. Value of decision making information based on simulation is also discussed.
B.03 Paper at an international scientific conference
COBISS.SI-ID: 6501907