Nature Inspired Computational Optimisation Methods for System Dynamics
MetadataShow full item record
This item's downloads: 1868 (view details)
This thesis aims to provide the state of the art "nature-inspired computing" (NIC) approaches and investigate their applications in system dynamics (SD) models. Specifically, we will focus on particle swarm optimisation (PSO), genetic algorithms (GAs) and cooperative coevolutionary approaches (CCEAs), and their application in the beer distribution game (BDG). Three pieces of work have been conducted in this thesis. Firstly, we focus on improving the performance of PSO. PSO is an intelligent random search algorithm, and the key to its success is to effectively balance between the exploration and the exploitation of the solution space. This thesis presents a new dynamic topology called "gradually increasing directed neighbourhoods (GIDN)". Each particle begins with a small number of connections and there are many small isolated swarms that improve the exploration ability. At each iteration, we gradually add a number of new connections between particles which improves the ability of exploitation. A series of experiments show that the PSO with GIDN performs much better than a number of the state of the art algorithms. Secondly, we are concerned with the applications of the PSO approaches to the BDG. The BDG offers a complex simulation environment involving multidimensional constrained pa- rameters. In order to obtain the optimal strategies for the BDG, we proposed the use of our PSO with GIDN. A number of the state of the art PSO algorithms and GAs have also been used as benchmarks. Two scenarios for the BDG are examined. In one scenario, all sectors use the same inventory management strategies while, in the other one, each sector has differ- ent strategies. The optimal strategies and their performance for both scenarios are examined. Furthermore, our PSO's performance is also investigated. Finally, we investigate the applications of CCEAs to the BDG. CCEAs are a flexible multi- population based framework. Each population can evolve independently and have their own goals. We have developed a coevolutionary framework for evolving strategies across this four- tier BDG. Our results identify the effects of two different management strategies on the supply chain performance. The first is where sectors are individually oriented, and the second is where sectors cooperate to achieve a common goal and are group oriented. We design two fitness approaches to reflect both management strategies. We have used this framework to identify the impact of these two strategies on the supply chain performance. Furthermore, two different demand patterns: a step input and uniformly distributed demand are also examined. A series of managerial insights have been derived from our extensive simulations.
This item is available under the Attribution-NonCommercial-NoDerivs 3.0 Ireland. No item may be reproduced for commercial purposes. Please refer to the publisher's URL where this is made available, or to notes contained in the item itself. Other terms may apply.
The following license files are associated with this item: