Elevator kinematics optimization : a novel meta-heuristic method

Name: Pasura Aungkulanon

keyword: Meta-heuristic

; Elevator kinematics optimization

; Non-constrained function

; Single objective model

; Multi-objective model

; Aggregate production planning

Abstract: The main objective of this research is to develop the sequential procedures for optimizing engineering problems. The novel algorithm based on Elevator Kinematics Optimization or EKO is conceptualized from elevator group control and traveling distance operated by elevator kinematics functions to maximize passenger satisfaction. Kinematics functions were classified into three conditions of A, B and C. These kinematic conditions can be categorized by the probability distribution (P-based) or the traveling distance (D-based). For the first condition, the maximum velocity and maximum acceleration would not be attained. For the second condition, the maximum acceleration would not be attained and the last one, the maximum velocity would be attained. After an operation of one among three elevator kinematics functions, the External Random Command for changing the destination from the passengers is also included to provide the satisfactory elevator system. There are six improvement procedures which consist of an initialization of elevator parameters, a preparation of the elevator memory, an improvement of a new floor from the elevator memory, an update of the elevator memory by D-based and P-based, an application of a random external command and a determination of stopping criterion. In this study, various engineering optimization problems were used to determine the performance of the EKO. They consisted of noisy response surface functions, single and multi-objective programming problems including a complex industrial problem of an aggregate production planning or APP. Response functions included the single peak of Parabolic function, the multi-peak of Camelback, Rastrigin and Shekel functions, the curved ridge of Rosenbrock and Styblinski functions and the multi-peak with curved ridge of Branin Goldstein-Price functions. Experimental results were compared in terms of maximum, minimum, mean and standard deviation. Previous well-known meta-heuristics were used to compare to D-based and P-based EKOs. They were bee colony (BA), shuffled frog leaping (SFLA), harmony search (HSA), firefly (FA), particle swarm optimization (PSO) algorithms. The hybridization among these methods and the novel method were also carried out for non-constrained, constrained machining and complex industrial problems. In non-constrained optimization problems of response functions, the results obtained from the novel ones were similar to those obtained from other well-known meta-heuristics. In constrained machining problems, both EKO and HSFLA showed the best performance relating to precision and robustness of the results. In aggregate production planning problems, the EKO outperformed. In conclusion, the EKO was the promising method in solving complex engineering problems. All three conditions and external random command were the main tools in finding the optimum. Condition A is more powerful for fine tuning. Condition B and C could produce reliable solutions in search space whereas the random external command assisted in escaping from local trap. The EKO requires a small number of parameters which consist of an elevator memory (EM), acceleration (a), maximal velocity (Vmax), probability for rejecting a command (PRC) and traveling height index (THI)

Thammasat University. Thammasat University Library

Address: BANGKOK

Email: preserv@tu.ac.th

Name: Pongchanun Luangpaiboon

Role: advisor

Created: 2015

Modified: 2021-10-27

Issued: 2021-10-27

วิทยานิพนธ์/Thesis

application/pdf

DOI: 10.14457/TU.the.2015.1548

eng

DegreeName: Doctor of Philosophy

Level: Doctoral Degree

Descipline: Engineering

Grantor: Thammasat University

©copyrights Thammasat University

ลำดับที่.	ชื่อแฟ้มข้อมูล	ขนาดแฟ้มข้อมูล	จำนวนเข้าถึง	วัน-เวลาเข้าถึงล่าสุด
1	4777pasura.pdf	11.27 MB

ใช้เวลา

0.023173 วินาที