Self-assembly for supply chains

Self-assembly is a natural construction process where components of a system spontaneously form into more complex aggregates when suitable environmental conditions are created. Self-assembly systems are remarkable in that the fabrica-tion of the complex structures are done with mechanisms that are s...

Full description

Saved in:
Bibliographic Details
Main Author: Yee, Gabriel Qi Ming
Other Authors: Ong Yew Soon
Format: Thesis-Master by Research
Language:English
Published: Nanyang Technological University 2020
Subjects:
Online Access:https://hdl.handle.net/10356/142760
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Nanyang Technological University
Language: English
Description
Summary:Self-assembly is a natural construction process where components of a system spontaneously form into more complex aggregates when suitable environmental conditions are created. Self-assembly systems are remarkable in that the fabrica-tion of the complex structures are done with mechanisms that are self-reproducing and maintaining, distributed, and are not restricted to having be synchronous. From the perspective of strategy development, such bottom-up behaviours are like the real-world process of systematically identifying and studying the key issues and reasons for a problem before matching it with a strategy to solve it. In a similar fashion, the real-world processes of specifying objectives, tasks, and principles are like the specifying of environmental condi-tions when designing self-assembly systems. These two behaviours exist as two extreme ends of strategy development causing the typical academic publication on strategy development to dichotomously adopt one. As a science that can bridge both approaches, the ability to self-assembly a strategy would present a superior approach to strategy development. In this thesis, the conceptualization and implementation of an algorithm that self-assembles a strategy is presented. The algorithm is applied to a supplier se-lection problem and benchmarked as a symbolic regression solver against tradi-tional Genetic Programming across five representative problems. Finally, the thesis is concluded with statements for potential extension.