Analysis of Nature-Inspired Algorithms for Long-Term Digital Preservation

Abstract

Digital preservation is a research area devoted to keeping digital assets preserved and usable for many years. Out of the many approaches to digital preservation, the present research article follows a new object-centered digital preservation paradigm where digital objects share part of the responsibility for preservation: they can move, replicate, and evolve to a higher-quality format inside a digital ecosystem. In the new framework, the behavior of digital objects needs to be modeled in order to obtain the best preservation strategy. Thus, digital objects are programmed with the mission of their own long-term self-preservation, which entails being accessible and reproducible by users at any time in the future regardless of frequent technological changes due to software and hardware upgrades. Three nature-inspired computational intelligence algorithms, based on the collective behavior of decentralized and self-organized systems, were selected for the modeling approach: multipopulation genetic algorithm, ant colony optimization, and a virus-based algorithm. TiM, a simulated environment for running distributed digital ecosystems, was used to perform the experiments. The results map the relation between the models and the expected object diversity obtained in short- and mid-term digital preservation scenarios. Comparing the results, the best performance corresponded to the multipopulation genetic algorithm. The article aims to be a first step in the digital self-preservation field. Building nature-inspired model behaviors is a good approach and opens the door to future tests with other AI-based methods