Optimizing Snowplowing Costs on Estonian Rural Roads - Comparative Approaches
Organisatsiooni nimi
Attention Merchants OÜ
Kokkuvõte
I would like to optimize snowplowing on snowplow routes in Estonia - time and distance should be miminized. The basic question is whether to choose urban or rural locations, let's for simplicity and low hanging fruit combination's sake take the key non-urban county in Estonia - Harjumaa.
Input data are
\tweather information for the entire year, 2021 - 2024 (public weather data - predictions + actual weather),
\troad information (OpenStreetMap https://www.openstreetmap.org/#map=8/58.613/25.024; Estonian Smart Road's Open Data https://imo.ut.ee/en/infrastructure/estonian-open-government-data/smart-roads-open-data/; "Riigiteataja: Tee seisunditase" https://www.riigiteataja.ee/akt/115072015013 - NB! the criticality of roads varies),
\tsnowplow fleet size for a given road network - say 25 plouging machines
\t
The problem could be solved with three approaches by three different teams:
\tOperations Research (OR) (see, https://journals.sagepub.com/doi/epdf/10.3141/2440-03, for a complete mixed integer programming (MIP) optimization solution for Approach 1 - hint, GPT is really useful for understanding the mathematical setup)
\tMachine Learning, supervised learning - Approach 2 could utilize this approach to solving the problem. Approach 2 should base the solution on the general approach taken by the OR MIP article.
\tMachine Learning, reinforcement learning - Approach 3 could utilize this approach to solving the problem. Approach 3 should base their solution on the general approach taken by the OR MIP article.
Input data are
\tweather information for the entire year, 2021 - 2024 (public weather data - predictions + actual weather),
\troad information (OpenStreetMap https://www.openstreetmap.org/#map=8/58.613/25.024; Estonian Smart Road's Open Data https://imo.ut.ee/en/infrastructure/estonian-open-government-data/smart-roads-open-data/; "Riigiteataja: Tee seisunditase" https://www.riigiteataja.ee/akt/115072015013 - NB! the criticality of roads varies),
\tsnowplow fleet size for a given road network - say 25 plouging machines
\t
The problem could be solved with three approaches by three different teams:
\tOperations Research (OR) (see, https://journals.sagepub.com/doi/epdf/10.3141/2440-03, for a complete mixed integer programming (MIP) optimization solution for Approach 1 - hint, GPT is really useful for understanding the mathematical setup)
\tMachine Learning, supervised learning - Approach 2 could utilize this approach to solving the problem. Approach 2 should base the solution on the general approach taken by the OR MIP article.
\tMachine Learning, reinforcement learning - Approach 3 could utilize this approach to solving the problem. Approach 3 should base their solution on the general approach taken by the OR MIP article.
Lõputöö kaitsmise aasta
2024-2025
Juhendaja
Kallol Roy, Jaan Übi
Suhtlemiskeel(ed)
eesti keel, inglise keel
Nõuded kandideerijale
Tase
Magister
Märksõnad
Kandideerimise kontakt
Nimi
Jaan Übi
Tel
E-mail