4683847_monash_130370.pdf (12.87 MB)
Download file

Misperception and its evolutionary value

Download (12.87 MB)
thesis
posted on 23.02.2017, 00:49 authored by Brumley, Lachlan
Misperception has a detrimental effect on an entity's perception of its environment, which can affect its decision-making abilities. To those with a perfect perception of an environment, the decisions and subsequent actions of a misperceiving entity may appear to be deficient in some manner. However, in some cases it is possible that an entity or its population ultimately benefits from the effects of misperception. This thesis aims to study how misperception occurs and better explain the conditions under which it may provide a benefit, evolutionary or otherwise. Misperception can have many possible causes, which may be divided into two main classes --- those caused by inadvertent flaws of the perceiving entity and those caused by the actions of other entities. Deliberate actions that are intended to cause an entity to misperceive are instances of Information Warfare attacks. This work produces a new generalised model that describes how both intentional and unintentional misperception affects entities during their decision-making process. These various sources of misperception are mapped into the Observation Orientation Decision Action (OODA) loop model, revealing how misperception affects the decision-making process in the short and long term. The internal process of the OODA loop is also examined and its Orientation step is expanded to detail how an entity's internal representation of the world is developed and affected by misperception, and also how this representation is used to make decisions. Historical case studies of some misperceptions are also mapped into this model, detailing the causes of such misperceptions. Previous research has identified some instances where misperception provides some benefit to misperceiving entities. This work also aims to identify what circumstances are required for entities to benefit from misperception and determine whether such a benefit can aid entities in an evolutionary environment. Artificial Life simulations are used to investigate environments where entities may misperceive. A benefit from misperception is demonstrated by a population of entities who evolve a significant level of misperception instead of correct perception. The results of these simulations reveal several different methods by which misperception may benefit both individual entities and their populations in different scenarios.

History

Campus location

Australia

Principal supervisor

Kevin Korb

Additional supervisor 1

Carlo Kopp

Year of Award

2014

Department, School or Centre

Clayton School of Information Technology

Course

Doctor of Philosophy

Degree Type

DOCTORATE

Faculty

Faculty of Information Technology