Inteligentni algoritem za optimiranje montažnih in strežnih sistemov ter procesov linijske proizvodnje

doktorsko delo

Hugo Zupan (Avtor), Niko Herakovič (Mentor)

Povzetek

Successful improvement of the competitiveness of enterprises depends to a large extent on the efficiency of assembly and handling systems and processes (AHSP). Their efficiency can be enhanced through various optimization methods, in particular in terms of the cost reduction, reduction of the throughput times, delivery times, increased utilization of equipment, etc. One of the most effective methods for optimizing such systems is optimization with on-line simulation. Such approach requires detailed research, study and analysis of all the building blocks and parameters needed to set up an expert system of on-line simulation of AHSP of the production line. Therefore, in the doctoral thesis, the development of intelligent algorithms and the use of them for the development of digital agents and expert systems of AHSP production line is discussed in detail. The expert system, developed in the doctoral thesis, in connection with the digital twin and digital agents, constantly monitors and continuously optimizes AHSP of the production line. In the doctoral thesis, an intelligent algorithm, called "flip and insert" is developed that can automatically suggest a very competitive schedule of orders, machines, etc. in a shorter time than well-established comparable algorithms. For the needs of validating the expert system with an intelligent algorithm, a real system of production line has been built in the laboratory environment. We combined the digital AHSP with the real system over the cloud, and thus set up all the necessary frameworks of the on-line simulation and thus develop an expert system that is in constant connection with the real system and is constantly monitoring and optimizing it. The methodology for intelligent algorithm, digital agents and digital twins provides a framework for their practical application in a real production environment.

Ključne besede

inteligentni algoritmi;optimiranje proizvodnih linij;sprotna simulacija;digitalni dvojčki;digitalni agenti;ekspertni sistemi;disertacije;

Podatki

Jezik:	Slovenski jezik
Leto izida:	2019
Tipologija:	2.08 - Doktorska disertacija
Organizacija:	UL FS - Fakulteta za strojništvo
Založnik:	[H. Zupan]
UDK:	658.527:004.021:004.942(043.3)
COBISS:	16717339
Št. ogledov:	964
Št. prenosov:	320
Ocena:	0 (0 glasov)
Metapodatki:

Ostali podatki

Sekundarni jezik:	Angleški jezik
Sekundarni naslov:	Intelligent algorithm for the optimization of assembly and handling systems and processes for in-line production
Sekundarni povzetek:	Povečanje konkurenčnosti podjetij je v veliki meri odvisno od učinkovitosti montažnih in strežnih sistemov ter procesov (MiSSP). Njihovo učinkovitost lahko povečamo z različnimi metodami optimizacije, predvsem z vidika zmanjševanja stroškov, skrajševanja pretočnih časov, dobavnih terminov, povečanja izkoriščenosti opreme itd. Ena najbolj učinkovitih metod optimiranja takšnih sistemov je optimiranje s sprotno simulacijo. Takšen pristop zahteva podrobne raziskave, študije in analizo vseh gradnikov ter parametrov, ki so potrebni, da se postavi ekspertni sistem sprotne oziroma "On-line" simulacije MiSSP linijske proizvodnje. V doktorski disertaciji je zato podrobno obravnavan razvoj inteligentnih algoritmov in uporaba le-teh za razvoj digitalnih agentov ter ekspertnih sistemov MiSSP linijske proizvodnje. Ekspertni sistem, razvit v okviru doktorske naloge, v povezavi z digitalnim dvojčkom in digitalnimi agenti nenehno nadzoruje in sproti optimira MiSSP linijske proizvodnje. Prav tako je v doktorskem delu razvit inteligentni algoritem, imenovan "premešaj in vstavi", ki npr. samodejno predlaga najboljše rešitve razporejanja naročil, strojev itd. v krajšem času kot uveljavljeni primerljivi algoritmi. Za potrebe validacije ekspertnega sistema z inteligentnim algoritmom je bil v laboratorijskem okolju zgrajen realni montažni in strežni sistem linijske proizvodnje. Digitalni MiSSP smo združili z realnim sistemom preko oblaka in s tem postavili vse potrebne okvire sprotne ali "On-line" simulacije in tako razvili ekspertni sistem, ki je v nenehni povezavi z realnim sistemom in ga sproti nadzoruje ter optimira. Postavljena metodologija zasnove inteligentnega algoritma, digitalnih agentov in digitalnih dvojčkov omogoča okvir za njihovo praktično uporabo v realnem proizvodnem okolju.
Sekundarne ključne besede:	dissertations;intelligent algorithms;optimization;production lines;on-line simulation;digital twins;digital agents;expert systems;Proizvodnja;Disertacije;Optimiranje;Simulacija;
Vrsta dela (COBISS):	Doktorsko delo/naloga
Študijski program:	0
Komentar na gradivo:	Univ. v Ljubljani, Fak. za strojništvo
Strani:	XXIV, [107] str., [7] str. pril.
ID:	11188447