Work

Loputtomien muodostelmien tuottaminen algoritmeilla

Year: 2021
Type: dissertation
Abstract: In this master's thesis, we explore how an infinite formation can be achieved in a computer program. We call an infinite formation a data structure that, when expressed in a coordinate system, extends indefinitely in at least one direction. An infinite formation, as the name suggests, comprises an infinite amount of data. Because we do not have infinite storage space, infinite formations have to be procedurally generated. They cannot be fully generated, however, as it is not possible to finish generating a structure that has no end. By dividing an infinite formation into parts that are finite in size, the parts can be generated and utilized. Therefore, in order to use an infinite formation, it has to be represented by an algorithm that can procedurally generate the formation in parts. We examine three ways of procedurally representing infinite formations: fractals, hash functions and noise functions. To evaluate the suitability of these methods, we compare their products with a theoretical, perfect infinite formation. An ideal infinite formation is deterministic, consists of unique parts, and all its parts are accessible in constant time. Our conclusion is that of the methods considered, noise functions produce infinite formations with the best qualities.
Author: Hirvonen, Jussi
Institution: Aalto University
Language: English
Open Access status: green
OpenAlex ID: W7112785397
Cites: 13
Cited by:
Related to: 10