A Systematic Study on Solving Aerospace Problems Using Metaheuristics

Complex engineering problems can be modelled as optimisation problems. For instance, optimising engines, materials, components, structure, aerodynamics, navigation, control, logistics, and planning is essential in aerospace. Metaheuristics are applied to solve these optimisation problems. The present paper presents a systematic study on applying metaheuristics in aerospace based on the literature. Relevant scientific repositories were consulted, and a structured methodology was used to filter the papers. Articles published until March 2022 associating metaheuristics and aerospace applications were selected. The most used algorithms and the most relevant hybridizations were identified. This work also analyses the main types of problems addressed in the aerospace context and which classes of algorithms are most used in each problem.

High-level hybridization of heuristics and metaheuristics to solve symmetric TSP: a comparative study

The Travelling Salesman Problem - TSP is one of the most explored problems in the scientific literature to solve real problems regarding the economy, transportation, and logistics, to cite a few cases. Adapting TSP to solve different problems has originated several variants of the optimization problem with more complex objectives and different restrictions. Metaheuristics have been used to solve the problem in polynomial time. Several studies have tried hybridising metaheuristics with specialised heuristics to improve the quality of the solutions. However, we have found no study to evaluate whether the searching mechanism of a particular metaheuristic is more adequate for exploring hybridization. This paper focuses on the solution of the classical TSP using high-level hybridisations, experimenting with eight metaheuristics and heuristics derived from k-OPT, SISR, and segment intersection search, resulting in twenty-four combinations. Some combinations allow more than one set of searching parameters. Problems with 50 to 280 cities are solved. Parameter tuning of the metaheuristics is not carried out, exploiting the different searching patterns of the eight metaheuristics instead. The solutions' quality is compared to those presented in the literature.

Attention-Driven Multi-Agent Reinforcement Learning: Enhancing Decisions with Expertise-Informed Tasks

In this paper, we introduce an alternative approach to enhancing Multi-Agent Reinforcement Learning (MARL) through the integration of domain knowledge and attention-based policy mechanisms. Our methodology focuses on the incorporation of domain-specific expertise into the learning process, which simplifies the development of collaborative behaviors. This approach aims to reduce the complexity and learning overhead typically associated with MARL by enabling agents to concentrate on essential aspects of complex tasks, thus optimizing the learning curve. The utilization of attention mechanisms plays a key role in our model. It allows for the effective processing of dynamic context data and nuanced agent interactions, leading to more refined decision-making. Applied in standard MARL scenarios, such as the Stanford Intelligent Systems Laboratory (SISL) Pursuit and Multi-Particle Environments (MPE) Simple Spread, our method has been shown to improve both learning efficiency and the effectiveness of collaborative behaviors. The results indicate that our attention-based approach can be a viable approach for improving the efficiency of MARL training process, integrating domain-specific knowledge at the action level.

Hybrid Genetic Algorithm and Mixed Integer Linear Programming for Flying Sidekick TSP

The increasing use of drones to perform various tasks has motivated an exponential growth of research aimed at optimizing the use of these means, benefiting both military and civilian applications, including logistics delivery. In this sense, the combined use of trucks and drones has been explored with great interest by Operations Research. This work presents mathematical formulations in Mixed Integer Linear Programming and proposes a hybrid Genetic Algorithm (HGenFS) for optimizing a variation of the Traveling Salesman Problem (TSP) called Flying Sidekick TSP (FSTSP), in which truck and drone cooperate. The results obtained confirmed that the adopted formulation for the exact solution is suitable for solving problems up to ten customers, and the HGenFS proved to be capable of finding optimal solutions for the FSTSP in a few seconds by incorporating specific heuristics and a local search phase.