Modeli za rešavanje problema lociranja resursa u sistemima povratne logistike

Abstract

Problem sa kojim se suočava današnje društvo je iznalaženje načina za tretman proizvoda koji su odbačeni od strane krajnjih korisnika (povratni tokovi proizvoda) u cilju redukcije otpada i očuvanja prirodnih resursa, a time i povećanja udela otpada koji se može ponovo koristiti, uz istovremeno smanjenje opterećenja deponijskog prostora. Rešenje problema upravljanja povratnim tokovima proizvoda, materijala i roba, pruža povratna logistika (eng. reverse logistics) jer ponovno korišćenje proizvoda, u bilo kom obliku donosi korist, kako u ekološkom, tako i u socioekonomskom pogledu. Podupiranje ove tvrdnje naučnim metodama je jedan od glavnih ciljeva povratne logistike. Efikasno i efektivno upravljanje povratnim tokovima proizvoda može doneti direktnu korist kompanijama kroz smanjenje sirovina u proizvodnji, stvaranje novih vrednosti kroz oporavak i obnavljanje vraćenih proizvoda, ili kroz smanjenje troškova odlaganja otpada na način koji je ekološki, društveno i zakonski prihvatljiv. Jasno je da finansijski uspeh sprovođenja aktivnosti povratne logistike određen, odnosno zavisi od operativne profitabilnosti uspostavljene logističke mreže. Kako bi maksimizirale vrednost korišćenih i vraćenih proizvoda, kompanije moraju da uspostave logističke strukture koje će upravljati povratnim tokovima robe na optimalan način. Iz tog razloga predmet ovog istraživanja je projektovanje mreža povratne logistike, istovremeno modelirajući probleme rutiranja vozila i lokacije objekata, odnosno jednovremeno posmatrajući i strateški i taktički/operativni nivo odlučivanja, koji kao takav donosi uštede na duži vremenski period, za razliku od nezavisnog rešavanja problema lokacije objekata i rutiranja vozila. U istraživanju je predstavljen matematički model mešovitog celobrojnog linearnog programiranja za definisanje optimalne mreže za reciklažu ambalažnog otpada, formulisan kao lokacijski-ruting model (eng. Location Routing Problem, LRP) i heuristički algoritam za njegovo rešavanje...

The problem facing society today, is finding ways to treat products discarded by th end users (i.g. reverse flows of products), in order to reduce waste and preserve natural resources, thereby increasing the proportion of waste that can be reduced, while reducing the amount of waste sent to landfills. Solution to this problem gives reverse logistics, since the re-use of the products, in any form, provides benefits both in the environmental and socio-economic terms. Supporting this statement with scientific methods is one of the main goals of reverse logistics. Efficient and effective management of reverse flows of products can bring direct benefits to companies through reduction in raw materials, creating new value through recovery and recovery of returned products, or by reducing the cost of waste disposal in a way that is environmentally, socially and legally acceptable. It is clear that the financial success of the reverse logistics activities implementation depends on the operating profitability of established logistics network. To maximize the value of used and returned products, companies need to establish a logistics structure that will manage the return flow of goods in an optimal way. Hence, the subject of this research is to propose one way to design reverse logistics network, by simultaneously modeling problems of vehicle routing and facility locations. That is, simultaneously defining strategic and tactical / operational level decisions, which, as such, brings savings for a longer period of time, as opposed to independent solving problems of facility location and vehicle routing. This research presents a mathematical model of mixed integer linear programming to define an optimal network for the recycling of packaging waste, defined as the location-routing model (eng. Location Routing Problem, LRP) and a heuristic algorithm to solve it. Solutions obtained by the proposed heuristic algorithm are close to optimal or even optimal. The proposed model is tested for a different system parameters in order to gain more insight into system performance...