A Genetic Algorithm for Multi-component Optimization Problems: The Case of the Travelling Thief Problem

Vieira, Daniel; Soares, Gustavo Luís; Vasconcelos, J.A.; Mendes, Maria Célia

doi:10.1007/978-3-319-55453-2_2

Cited by 9 publications

(4 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Additionally, the authors investigated a TTP-specific local search algorithm. A Genetic Algorithm was used in [14]. Authors solve the overall problem instead of solving the subproblems separately.…”

Section: Related Workmentioning

confidence: 99%

A Specialized Evolutionary Approach to the bi-objective Travelling Thief Problem

Laszczyk¹,

Myszkowski²

2019

Proceedings of the 2019 Federated Conference on Computer Science and Information Systems

View full text Add to dashboard Cite

In the recent years, it has been shown that real world-problems are often comprised of two, interdependent subproblems. Often, solving them independently does not lead to the solution to the entire problem. In this article, a Travelling Thief Problem is considered, which combines a Travelling Salesman Problem with a Knapsack Problem. A Non-Dominated Sorting Genetic Algorithm II (NSGA-II) is investigated, along with its recent modification-a Non-Dominated Tournament Genetic Algorithm (NTGA). Each method is investigated in two configurations. One, with generic representation, and genetic operators. The other, specialized to the given problem, to show how the specialization of genetic operators leads to better results. The impact of the modifications introduced by NTGA is verified. A set of Quality Measures is used to verify the convergence, and diversity of the resulting PF approximations, and efficiency of the method. A set of experiments is carried out. It is shown that both methods work almost the same when generic representation is used. However, NTGA outperforms classical NSGA-II in the specialized results.

show abstract

Section: Related Workmentioning

confidence: 99%

A Specialized Evolutionary Approach to the bi-objective Travelling Thief Problem

Laszczyk¹,

Myszkowski²

2019

Proceedings of the 2019 Federated Conference on Computer Science and Information Systems

View full text Add to dashboard Cite

show abstract

“…Real-world problems, like supply chains, are characterised by interactions, where subproblems features are linked [2]. Hence, an optimal solution for each subproblem might not guarantee optimal overall solutions [12].…”

Section: Introductionmentioning

confidence: 99%

Facility location problem and permutation flow shop scheduling problem

Ogunsemi

McCall

Kern

et al. 2022

Proceedings of the Genetic and Evolutionary Computation Conference Companion

View full text Add to dashboard Cite

There is a growing literature spanning several research communities that studies multiple optimisation problems whose solutions interact, thereby leading researchers to consider suitable approaches to joint solution. Real-world problems, like supply chain, are systems characterised by such interactions. A decision made at one point of the supply chain could have significant consequence that ripples through linked production and transportation systems. Such interactions would require complex algorithmic designs. This paper, therefore, investigates the linkages between a facility location and permutation flow shop scheduling problems of a distributed manufacturing system with identical factory (FLPPFSP). We formulate a novel mathematical model from a linked optimisation perspective with objectives of minimising facility cost and makespan. We present three algorithmic approaches in tackling FLPPFSP; Nondominated Sorting Genetic Algorithm for Linked Problem (NS-GALP), Multi-Criteria Ranking Genetic Algorithm for Linked Problem (MCRGALP), and Sequential approach. To understand FLPPFSP linkages, we conduct a pre-assessment by randomly generating 10000 solution pairs on all combined problem instances and compute their respective correlation coefficients. Finally, we conduct experiments to compare results obtained by the selected algorithmic methods on 620 combined problem instances. Empirical results demonstrate that NSGALP outperforms the other two methods based on relative hypervolume, hypervolume and epsilon metrics. CCS CONCEPTS• Applied computing → Supply chain management.

show abstract

“…Real-world problems, like service chains, are systems characterised by such interdependency, where some features of the subcomponents of the problem are linked [2]. In such linkages, an optimal solution for individual operational components might not guarantee an optimal solution for the overall problem [3]. An example is the integration between job assignment (JAP) and travelling salesman problem (TSP) of a service chain system where service personnel perform tasks at different locations.…”

Section: Introductionmentioning

confidence: 99%

Job Assignment Problem and Traveling Salesman Problem: A Linked Optimisation Problem

Ogunsemi

McCall

Kern

et al. 2022

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Linked decision-making in service management systems has attracted strong adoption of optimisation algorithms. However, most of these algorithms do not incorporate the complexity associated with interacting decision-making systems. This paper, therefore, investigates the linkages between two classical problems: job assignment problem and travelling salesman problem (JAPTSP) of a service chain system where service personnel perform tasks at different locations. We formulate a novel mathematical model from a linked optimisation perspective with objectives to minimise job cost and total travel distance simultaneously. We present three algorithmic approaches to tackling the JAPTSP: Nondominated Sorting Genetic Algorithm for Linked Problem (NSGALP), Multi-Criteria Ranking Genetic Algorithm for Linked Problem (MCR-GALP), and Sequential approach. We evaluate the performance of the three algorithmic approaches on a combination of JAP and TSP benchmark instances. Results show that selecting an appropriate algorithmic approach is highly driven by specific considerations, including multiobjective base performance metrics, computation time, problem correlation and qualitative analysis from a service chain perspective.

show abstract

A Genetic Algorithm for Multi-component Optimization Problems: The Case of the Travelling Thief Problem

Cited by 9 publications

References 9 publications

A Specialized Evolutionary Approach to the bi-objective Travelling Thief Problem

A Specialized Evolutionary Approach to the bi-objective Travelling Thief Problem

Facility location problem and permutation flow shop scheduling problem

Job Assignment Problem and Traveling Salesman Problem: A Linked Optimisation Problem

Contact Info

Product

Resources

About