Integrisano planiranje i terminiranje tehnoloških procesa u dinamičkim uslovima - pregled stanja u oblasti istraživanja

Abstract

U ovom radu je dat detaljan pregled stanja u oblasti istraživanja jedne od funkcija inteligentnih tehnoloških sistema (ITS) - integrisano planiranje i terminiranje tehnoloških procesa u dinamičkim uslovima (DIPPS). U tom smislu, datje opis DIPPSproblema, razmatrani su kriterijumi na osnovu kojih se vrši odabir optimalanog plana terminiranja, definisane su usvojene pretpostavke i predstavljen je matematički model ovog problema. Takođe, detaljno su razmatrani i sledeći poremećajni faktori koji se mogu javiti u okviru tehnoloških sistema: (i) prestanak rada mašine alatke, (ii) dolazak novog dela u sistem i (iii) otkaz obrade dela. Analizirani su pristupi za rešavanje DIPPS problema bazirani na multiagentnim sistemima, kao i pristupi bazirani na algoritmima. Kada su u pitanju pristupi bazirani na algoritmima, fokus u ovom radu je na biološki inspirisanim algoritmima optimizacije i to: evolucionim algoritmima, algoritmima baziranim na inteligenciji roja, kao i hibridnim pristupima. Kritičkom analizom stanja u ovoj oblasti istraživanja može se zaključiti da biološki inspirisane tehnike veštačke inteligencije imaju veliki potencijal u optimizaciji pomenute funkcije ITS-a.

This paper gives a detailed state-of-the art in the research area o f the important function o f Intelligent Manufacturing Systems (IMS) - integrated process planning and scheduling o f manufacturing systems in dynamic environment (DIPPS). Referring to this, description o f the DIPPS problem is given, the criteria on the basis o f which the optimal rescheduling plan are formulated and considered, the adopted assumptions are defined and the mathematical model o f this problem is presented. Furthermore, the disturbances that occur in manufacturing systems are considered in detail: (i) machine breakdown, (ii) arrival of a new job and (iii) job cancellation. Approaches for solving DIPPS problems based on multiagent systems as well as approaches based on algorithms are analyzed. When it comes to approaches based on algorithms, the focus of this paper is on biologically inspired optimization algorithms: evolutionary algorithms, swarm intelligence based algorithms as well as hybrid approaches. The critical analysis within this research area is shown in order to conclude that biologically inspired artificial intelligence techniques have great potential in optimizing the considered IMS function.