Optimiranje omjera reaktivnog i proaktivnog održavanja opreme korištenjem Markovljevog procesa odlučivanja

master's thesis

• Title (english): Optimizing the ratio of reactive and proactive maintenance of equipment using Markov decision process
• Author: Bruno Jerić
• Mentor: Mirko Randić (mentor)
• Committee member: Mirko Randić (predsjednik povjerenstva)
• Committee member: Bruno Blašković (član povjerenstva)
• Committee member: Luka Humski (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2023-07-05, Croatia

Abstract

U radu je ukratko prikazana teorija Markovljevih lanaca, Markovljevih procesa odlučivanja
i polu-Markovljevih procesa odlučivanja. Modeliran je i opisan problem iz stvarnog svijeta
koji se odnosi na dinamiku održavanja telekomunikacijskih linija. U model su ugrađena tri
stanja linije sa pripadajućimmatricama prijelaza i nagradama. Za slučaj SMDP-a modelirana
su i vremena koja predstavljaju ispravan rad linije kao i kontinuirana dobit tijekom vremena
kada je linija ispravna.
Implementirani su algoritmi iteracije politike u programskom jeziku Python. Za slučaj kada
je linija modelirana MDP-om određene su dvije politike - deterministička i stohastička. A za
SMDP, određena je optimalna politika koja u obzir uzima vremena i kontinuirane nagrade

Abstract (english)

The paper covers the theory of Markov chains, Markov decision processes, and Semi-Markov decision processes. A real-world problem is modeled and described. A telecommunications line is modeled with corresponding transition matrices and rewards in an MDP. Additionally, for the case of SMDP, times representing the correct operation of the line and continuous profit when the line is functioning properly are modeled.
Policy iteration algorithms were implemented in the Python programming language. For the case where the line is modeled as an MDP, two policies were found: deterministic and stochastic. For the SMDP case, an optimal policy considering times and continuous rewards was found.

• Keywords: Markovljevi lanci; Markovljev proces nagrađivanja; Markovljev proces odlučivanja; Polu-Markovljev proces odlučivanja; optimizacija troškova održavanja; reaktivno održavanje; proaktivno održavanje
• Keywords (english): Markov chains; Markov reward process; Markov decision process; Semi-Markov decision process; optimization of maintenance costs; reactive maintenance; proactive maintenance
• Language: croatian
• URN:NBN: urn:nbn:hr:168:001920
• 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: 2023-07-13 10:09:15