Optimiranje politike održavanja opreme korištenjem Markovljevih procesa odlučivanja

master's thesis

• Title (english): Optimization of equipment maintenance policy by using Markov decision processes
• Author: Luka Štroliga
• Mentor: Mirko Randić (mentor)
• Committee member: Mirko Randić (predsjednik povjerenstva)
• Committee member: Frano Škopljanac-Mačina (član povjerenstva)
• Committee member: Luka Humski (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2024-07-16, Croatia

Abstract

Dobro izabrana politika održavanja telekomunikacijske opreme ključna je za smanjenje
troškova i povećanje dostupnosti sustava, a uz to i direktno utječe na zadovoljstvo korisnika.
U ovom radu istražuje se mogućnost pronalaska idealne politike održavanja
telekomunikacijskih linija pomoću polu-Markovljevih procesa odlučivanja (SMDP).
Markovljev proces odlučivanja omogućava modeliranje sustava uzimajući u obzir trenutna
stanja sustava te akcije koje se mogu poduzeti u tim stanjima. SMDP model još dodaje i
vremenski aspekt tako da preciznije opisuje proces odlučivanja. Specificirana su dva modela
održavanja, a algoritam za rješavanje oba modela implementiran je u programskom jeziku
Python. Nakon provedene analize možemo reći da je pronađena idealna politika održavanja
za oba modela te se uz pomoć dobivenih rezultata može značajno smanjiti operativne
troškove održavanja uz povećanje raspoloživosti sustava.

Abstract (english)

A well-chosen telecommunications equipment maintenance policy is essential for reducing
costs and increasing system availability, and it also directly affects user satisfaction. In this
paper, the possibility of finding an ideal policy for the maintenance of telecommunication
lines using semi-Markov decision processes (SMDP) is assessed. The Markov decision
process enables system modeling by considering the current state of the system and the
actions that can be taken in those states. The SMDP model also adds a time aspect so that it
more precisely describes the decision-making process. Two maintenance models were
specified, and the algorithm for solving both models was implemented in the Python
programming language. After the analysis, we can say that we have found an ideal
maintenance policy for both models, and with the help of the obtained results, operational
maintenance costs can be significantly reduced while increasing system availability.

• Keywords: Markovljevi lanci; Markovljev proces nagrađivanja; Markovljevi procesi odlučivanja; Polu-Markovljev procesi odlučivanja
• Keywords (english): Markov chains; Markov reward process; Markov decision process; Sermi-Markov decision process
• Language: croatian
• URN:NBN: urn:nbn:hr:168:174862
• Study programme: Title: Computing; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra inženjer/inženjerka računarstva (magistar/magistra inženjer/inženjerka računarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2024-10-12 22:18:44