Publication: Risk-based drone routing using a hybrid swarm intelligence algorithm
Abstract
Son yıllardaki teknolojik gelişmelerle drone kullanımı artmıştır. Günümüzde farklı alanlarda görev yapan çeşitli drone’lar bulunmakta ve askeri harekat, gözetleme ve keşif, afet yönetimi, çevre modelleme veya lojistik gibi birçok alanda kullanımı yaygınlaşmaya devam etmektedir. Bu araçlar kullanımı boyunca kısıtlama alanı, aşırı hava koşulları (sıcak hava, rüzgar vb.), saldırılar veya coğrafi koşullar gibi bazı tehditlerle karşılaşabilmektedir. Bu tehditler altında optimum rotanın belirlenmesi oldukça önemlidir. Özellikle sürü zekası algoritmaları, bu tür karmaşık optimizasyon problemlerini çözmek için etkili çözümler sunar. Sürü Zekası Algoritmaları, doğadaki canlıların davranışlarından ilham alır ve bu davranışları yapay sistemlere entegre eder. Bu tezde teslimat ve lojistik sektöründeki kullanımı gittikçe artmakta olan drone rotalama problemi ele alınmıştır. Günümüzde pek çok gelişmiş firmanın teslimatlarını iyileştirebilmek ve süreçlerini optimize edebilmek için drone ile teslimat uygulamaları üzerine çalıştığı bilinmektedir. Bu çalışmada, gelecekte uygulanabilirliği yüksek, birden fazla teslimat yapabilen tek bir drone için risk-tabanlı bir rota optimizasyonu modeli geliştirilmiştir. Alışılagelmiş rota problemlerinin mesafe minimizasyonu amacına ek olarak, drone'un seyahati sırasında maruz kalacağı risk faktörünü de hesaba katarak rotanın minimum risk içerecek şekilde optimize edilmesi amaçlanmıştır. Problemin çözümünde, Harmony Search Algoritması (HS) ve Yapay Arı Kolonisi Algoritması (ABC) birleştirilerek, 2-Opt'lu Armoni Tabanlı Yapay Arı Kolonisi (HABC-2O) yaklaşımı uygulanmıştır. HABC-2O, klasik ABC ve Karınca Kolonisi Algoritması (ACO) ile karşılaştırılmıştır.
With technological developments in recent years, the use of drones has increased. Today, there are various drones operating in different fields, and their use continues to become widespread in many areas such as military operations, surveillance and reconnaissance, disaster management, environmental modeling or logistics. These vehicles may encounter some threats during their use, such as restriction area, extreme weather conditions (hot weather, wind, etc.), attacks or geographical conditions. Determining the optimum routeunder these threats is fundamental. Swarm intelligence algorithms, in particular, offer effective solutions to solve such complex optimization problems. Swarm Intelligence Algorithms are inspired by the behaviors of living creatures in nature and integrate these behaviors into artificial systems. In this thesis, the drone routing problem, which is increasingly used in the transportation and logistics sector, is discussed. Today, it is known that many developed companies are working on drone delivery applications in order to improve their deliveries and optimize their processes. In this study, a risk-based route optimization model is developed for a single drone with high future applicability and ability to make multiple deliveries. In addition to the distance minimization objective of the conventional routing problems, it is aimed to optimize the route to include minimum risk, considering the risk factor that the drone can be exposed to during its journey. In solving the problem, the 2-Opt Harmony Based Artificial Bee Colony (HABC-2O) approach is employed by combining the Harmony Search Algorithm (HS) and the Artificial Bee Colony Algorithm (ABC). HABC-2O has been compared to classical ABC and Ant Colony Algorithm (ACO).
With technological developments in recent years, the use of drones has increased. Today, there are various drones operating in different fields, and their use continues to become widespread in many areas such as military operations, surveillance and reconnaissance, disaster management, environmental modeling or logistics. These vehicles may encounter some threats during their use, such as restriction area, extreme weather conditions (hot weather, wind, etc.), attacks or geographical conditions. Determining the optimum routeunder these threats is fundamental. Swarm intelligence algorithms, in particular, offer effective solutions to solve such complex optimization problems. Swarm Intelligence Algorithms are inspired by the behaviors of living creatures in nature and integrate these behaviors into artificial systems. In this thesis, the drone routing problem, which is increasingly used in the transportation and logistics sector, is discussed. Today, it is known that many developed companies are working on drone delivery applications in order to improve their deliveries and optimize their processes. In this study, a risk-based route optimization model is developed for a single drone with high future applicability and ability to make multiple deliveries. In addition to the distance minimization objective of the conventional routing problems, it is aimed to optimize the route to include minimum risk, considering the risk factor that the drone can be exposed to during its journey. In solving the problem, the 2-Opt Harmony Based Artificial Bee Colony (HABC-2O) approach is employed by combining the Harmony Search Algorithm (HS) and the Artificial Bee Colony Algorithm (ABC). HABC-2O has been compared to classical ABC and Ant Colony Algorithm (ACO).