Publication detail

Minimisation of Network Covering Services with Predefined Centres

ŠEDA, M. ŠEDA, P.

Czech title

Minimalizace sítě servisních středisek s předdefinovanými centry

English title

Minimisation of Network Covering Services with Predefined Centres

Type

conference paper

Language

en

Original abstract

In this paper, we deal with a special version of the set covering problem, which consists in finding the minimum number of service centres providing specialized services for all customers (or larger units, such as urban areas) within a reasonable distance given by a threshold. If a suitable threshold is found that makes it possible to determine a feasible solution of the task, the task is transformed into a general set covering problem. In order to reflect the importance of the centers, we assign weights to them and, if some centers must be contained in the result, we can either add columns in the reachability matrix with link to these centres or add special constraints in the mathematical model. However, this is of a combinatorial nature and, because it belongs to the class of NP-hard problems, for a large instance of the problem, it cannot be used to and the optimal solution in a reasonable amount of time. In the paper, we present a solution that uses two heuristic methods: genetic algorithm and tabu search.

Czech abstract

V tomto příspěvku se zabýváme speciální verzí problému pokrytí, který spočívá v nalezení minimálního počtu servisních středisek zajišťujících specializované služby pro všechny zákazníky (nebo větší jednotky, jako jsou městké části) v přiměřené vzdálenosti do dané horní meze. Jestliže lze najít prahovou hodnotu, pro niž lze najít přípustné řešení, problém se nabývá obecné podoby. Avšak vzhledem ke kombinatorické povaze problému a jeho přískušnosti do třídy NP-těžkých problémů pro velké instance nelze najít otpimální řešení v dostupném čase. V příspěvku je prezentováno řešení, které využívá dvě heuristické metody: genetický algoritmus a zakázané prohledávání.

English abstract

In this paper, we deal with a special version of the set covering problem, which consists in finding the minimum number of service centres providing specialized services for all customers (or larger units, such as urban areas) within a reasonable distance given by a threshold. If a suitable threshold is found that makes it possible to determine a feasible solution of the task, the task is transformed into a general set covering problem. In order to reflect the importance of the centers, we assign weights to them and, if some centers must be contained in the result, we can either add columns in the reachability matrix with link to these centres or add special constraints in the mathematical model. However, this is of a combinatorial nature and, because it belongs to the class of NP-hard problems, for a large instance of the problem, it cannot be used to and the optimal solution in a reasonable amount of time. In the paper, we present a solution that uses two heuristic methods: genetic algorithm and tabu search.

Keywords in Czech

problém pokrytí, problém bez zohlednění vah, práh, matice dostupnosti, genetický algoritmus, opravný operátor

Keywords in English

Set covering, unicost problem, threshold, reachability matrix, genetic algorithm, repair operator

Released

22.05.2018

Publisher

European Council for Modelling and Simulation

Location

Wilhelmshaven (Germany)

ISBN

978-0-9932440-6-3

ISSN

2522-2414

Book

Proceedings of the 32nd European Conference on Modelling and Simulation ECMS 2018

Volume

2018

Number

1

Pages from–to

282–287

Pages count

6

BIBTEX


@inproceedings{BUT149158,
  author="Miloš {Šeda} and Pavel {Šeda},
  title="Minimisation of Network Covering Services with Predefined Centres",
  booktitle="Proceedings of the 32nd European Conference on Modelling and Simulation ECMS 2018",
  year="2018",
  volume="2018",
  number="1",
  month="May",
  pages="282--287",
  publisher="European Council for Modelling and Simulation",
  address="Wilhelmshaven (Germany)",
  isbn="978-0-9932440-6-3",
  issn="2522-2414"
}