ponedeljek, 16. marec 2026 Daniil Baldouski: Towards a closed-loop wood value chain: multi-objective optimization and real-world benchmarking

V ponedeljek, 16. marca 2026, bo ob 16:00 uri izvedeno 
predavanje v okviru PONEDELJKOVEGA SEMINARJA RAČUNALNIŠTVA IN INFORMATIKE 
Oddelkov za Informacijske znanosti in tehnologije UP FAMNIT in UP IAM.

ČAS/PROSTOR: 16. marec 2026 ob 16.00 v FAMNIT-VP3.

-----------------------------------------------
PREDAVATELJ: Daniil BALDOUSKI
-----------------------------------------------

Daniil Baldouski is a fourth-year PhD student of Computer Science at UP FAMNIT.

-----------------------------------------------------------------------------------------------------------------------------------------------------
NASLOV: Towards a closed-loop wood value chain: multi-objective optimization and real-world benchmarking
-----------------------------------------------------------------------------------------------------------------------------------------------------

POVZETEK: 

This work addresses the problem of designing closed-loop supply chain (CLSC) networks where production residuals and end-of-life products are reintegrated into the production cycle. A multi-objective mixed-integer linear programming (MILP) model is proposed that simultaneously minimizes economic costs and environmental impact. The model supports multiple echelons, multiple resource commodities, and resource transformation through pre-defined recipes, enabling the reuse of production residuals within the network wherever possible. A systematic benchmark generation methodology for CLSC optimization models is also introduced. The methodology supports controlled variation of network spatial structure through multiple location sampling methods and supply chain topology through alternative recipe scenarios. Real-world data from the Slovenian wood value chain is used to generate a benchmark dataset across different scales and configurations. Computational experiments show that small and medium instances are solved to optimality within reasonable time limits, while large instances present computational challenges depending on spatial configuration and supply chain complexity. The results also reveal an asymmetric trade-off between cost and environmental impact under different carbon pricing scenarios, where modest cost increases can lead to significant environmental benefits.

Seminar bo potekal v angleškem jeziku v predavalnici FAMNIT-VP3 s pričetkom ob 16:00 uri.

Vabljeni!