Perbandingan antara metoded coupling dan decoupling menggunakan collision based search dan flow-annotation replanning dalam algoritma multi agen=Comparison between coupling and decoupling method using collision based search and flow-annotation replanning in multi-agent pathfinding algorithm

Ngadin, Jasson (2020) Perbandingan antara metoded coupling dan decoupling menggunakan collision based search dan flow-annotation replanning dalam algoritma multi agen=Comparison between coupling and decoupling method using collision based search and flow-annotation replanning in multi-agent pathfinding algorithm. Bachelor thesis, Universitas Pelita Harapan.

	Text (Title) Title.pdf Restricted to Registered users only Available under License Creative Commons Attribution Non-commercial Share Alike. Download (3MB)
Preview	Text (Abstract) Abstract.pdf Available under License Creative Commons Attribution Non-commercial Share Alike. Download (45kB) | Preview
Preview	Text (ToC) ToC.pdf Available under License Creative Commons Attribution Non-commercial Share Alike. Download (189kB) | Preview
Preview	Text (Chapter1) Chapter1.pdf Available under License Creative Commons Attribution Non-commercial Share Alike. Download (166kB) | Preview
	Text (Chapter2) Chapter2.pdf Restricted to Registered users only Available under License Creative Commons Attribution Non-commercial Share Alike. Download (466kB)
	Text (Chapter3) Chapter3.pdf Restricted to Registered users only Available under License Creative Commons Attribution Non-commercial Share Alike. Download (974kB)
	Text (Chapter4) Chapter4.pdf Restricted to Registered users only Available under License Creative Commons Attribution Non-commercial Share Alike. Download (529kB)
	Text (Chapter5) Chapter5.pdf Restricted to Registered users only Available under License Creative Commons Attribution Non-commercial Share Alike. Download (148kB)
Preview	Text (Bibliography) Bibliography.pdf Available under License Creative Commons Attribution Non-commercial Share Alike. Download (43kB) | Preview
	Text (Appendices) Appendices.pdf Restricted to Repository staff only Available under License Creative Commons Attribution Non-commercial Share Alike. Download (283kB)

Abstract

Pendekatan utama yang diapakai untuk mengatasi sistem pencarian jalur multi-agen adalah metode coupling dan decoupling. Secara Teori metode coupling memberikan jalur optimal dan memberikan solusi, tetapi mempunyai waktu komputasi yang lama. Sedangkan metode decoupling lebih praktis dan memberikan hasil dengan cepat, walaupun tidak optimal, dan terkadang tidak memberi solusi. Eksperimen dilakukan menggunakan simulasi, dengan menggunakan labirin dengan ukuran 51x37, dengan 3 konfigurasi labirin yang berbeda. Dimana Conflict Based Search algorithm dan Flow Annotated Replanning Algorithm akan dilaksanakan. Eksperimen yang dilakukan secara total ada 75, proses testing melibatkan 2 algoritma untuk setiap labirin, dimulai dari 2 agen, sampai dengan 30 agen pada 2 labirin pertama, dan 15 pada labirin ketiga. Rata-rata tingkat pertumbahan pada labirin pertama adalaah 0.84% untuk Algoritma FAR, dan 10.4% untuk algoritma CBS, untuk rata-rata tingkat pertumbuhan pada memori pada labirin pertama adalah 0.25% untuk FAR, dan 2.85% untuk CBS. Rata-rata tingkat pertumbahan pada labirin kedua adalah 0.58 % untuk Algoritma FAR, dan 6.7% untuk algoritma CBS, untuk rata-rata tingkat pertumbuhan pada memori pada labirin kedua adalah 0.1% untuk FAR, dan 2.61% untuk CBS. Rata-rata tingkat pertumbahan pada labirin ketiga adalah 11.47% untuk Algoritma FAR, dan 18.48% untuk algoritma CBS, untuk rata-rata tingkat pertumbuhan pada memori pada labirin ketiga adalah 1.11% untuk FAR, dan 1.5% untuk CBS. / The main approaches when it comes to handling a multi agent pathfinding system, is through the method of coupling and decoupling. In theory, the coupling method gives and optimal solution and will always give a solution if there ever is one, but it is computationally intensive. While the decoupling gives a faster answer and less computation, it is also more practical and scalable to use, while not being the most optimal, and sometimes does not give a solution. The experiment is done though a simulation, which is in a 51x37 grid format, with 3 different maze configurations, it will simulate two algorithms within the maze, which is the Conflict Based Search algorithm, and the Flow Annotated Replanning Algorithm. In total, 75 experiments were done, the testing involves 2 algorithms for each maze, starting from 2 agents and reaching a cap of 30 agents for the first two maze, and 15 for the third maze. The average increase in time on the first maze is 0.84% for the FAR algorithm, and it is 10.4% for the CBS algorithm, while having an average memory growth rate 0.25% for FAR and 2.85% for CBS. The average increase in time on the second maze is 0.58% for the FAR algorithm, and it is 6.7% for the CBS algorithm, while having an average memory growth rate 0.1% for FAR and 2.61% for CBS. The average increase in time on the Third maze is 11.47% for the FAR algorithm, while the average increase in time is 18.48% for the CBS algorithm, while having an average memory growth rate 1.11% for FAR and 1.5% for CBS.

Actions (login required)

View Item