A modified micro-genetic algorithm for robust emergency system designing

Pesko, Stefan; Borcinova, Zuzana

doi:10.1109/carpathiancc.2019.8765916

Cited by 1 publication

(1 citation statement)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Such a finding may help in the determination of promising initial solutions for a searching process based on the neighborhood search, or it can provide metaheuristics imitating evolutionary processes or social behavior of species with a set of initial problem solutions. For example, genetic algorithms [20,21] start with a population of problem solutions, which should differ as much as possible. A scatter search algorithm [22] maintains two parts of a reference set of solutions, where the first part contains a collection of good solutions and the other one represents a maximally diversified set of solutions.…”

Section: Introductionmentioning

confidence: 99%

Kit of Uniformly Deployed Sets for p-Location Problems

2023

View full text Add to dashboard Cite

This paper deals with p-location problem solving processes based on a decomposition, which separates the creation of a uniformly deployed set of p-location problems from the solution of the p-location problem for that specific instance. The research presented in this paper is focused on methods of construction of uniformly deployed sets of solutions and the examination of their impact on the efficiency of subsequent optimization algorithms. The approaches to the construction are used for the constitution of predetermined families of uniformly deployed sets of p-location problem solutions, which have standard sizes. We introduce two methods of uniformly deployed set construction: the first one is based on composition, followed by an enlargement process; and the second one makes use of voltage graphs. The construction approaches are completed by an algorithm, which adjusts the set of solutions to the sizes of a solved instance. The influence of a set construction approach on solving process efficiency is studied on real-world benchmarks, which include both the p-median objective function and the generalized disutility function. The solving process is performed alternatively using the swap or path-relinking based methods. Results of the computational study obtained by all combinations of the mentioned approaches are presented and evaluated in the concluding part of the paper to make the studied characteristics visible.

show abstract

Section: Introductionmentioning

confidence: 99%