STUDENT Seminar

 

PROGRAM

---

Predavanja možete pratiti i online putem MITEAM stranice Studentskog seminara:
https://miteam.mi.sanu.ac.rs/asset/F37EF2gbfK8dWqxpw

---

Plan rada Studentskog seminara za MAJ 2026.

Petak, 22.05.2026. u 12:15, On-line
Marko Milenković, ETH, Cirih
KAKO PREVASPITATI TRIANGULACIJE SAT-OM
U ovom predavanju predstavićemo pobednički projekat tima ETH Flippers u juniorskoj kategoriji CG:SHOP 2026 izazova. Zadatak je pronaći centralnu triangulaciju za dati skup triangulacija nad istim skupom tačaka, tako da ukupan broj paralelnih flip-koraka do nje bude minimalan.
Objasnićemo osnovne geometrijske pojmove, SAT formulaciju problema i upotrebu XOR klauza za modelovanje promena ivica kroz vreme. Za velike instance, gde egzaktni pristup postaje preskup, prikazaćemo heuristike zasnovane na lokalnoj pretrazi i nezavisnim skupovima flipova. Predavanje pokazuje kako se računska geometrija, SAT rešavači i takmičarska intuicija mogu spojiti u efikasan pristup teškom optimizacionom problemu.

---

Petak, 22.05.2026. u 12:15, On-line
Maša Lazić i Ines Pavlović, Matematički fakultet, Univerzitet u Beogradu
OPTIMIZING MULTI-CLASS EV CHARGER PLACEMENT WITH GREEDY HEURISTICS
We study a capacitated facility location problem in which each facility has a limited capacity for serving customers. Customers may receive supply from multiple facilities, while some pairs of customers are mutually incompatible and cannot be assigned to the same facility due to logistical, operational, or other practical constraints.
The proposed model is applied to the planning of electric vehicle charging stations in Doha. Customers represent spatially distributed demand points, while facilities correspond to existing fuel stations that can potentially host EV chargers. The instances are generated using real-world geospatial population data and infrastructure data. This application is motivated by the high cost of EV charging infrastructure, since installing redundant chargers may lead to significant unnecessary investments. For example, ten unnecessary DC fast chargers could require an additional investment of approximately 200,000 to 600,000 USD, excluding annual maintenance costs.
The objective is to reduce redundant infrastructure while maintaining good demand coverage and respecting capacity and compatibility constraints. To solve the problem, we compare several metaheuristic techniques with respect to solution quality and execution time.

---

Predavanja su namenjena širokom krugu slušalaca. Održavaju se petkom sa početkom u 12:00 sati u sali 301f na trećem spratu zgrade Matematičkog instituta SANU, Knez Mihailova 36.

---