On the properties of artificial development and its use in evolvable hardware

Abstract

Abstract-The design of a new biologically inspired artificial developmental system is described in this paper. In general, developmental systems converge slower and are more computationally expensive than direct evolution. However, the performance trends of development indicate that the full benefit of development will arise with larger and more complex problems that exhibit some sort of regularity in their structure: thus, the aim is to evolve larger electronic systems through the modularity allowed by development. The hope is that the proposed artificial developmental system will exhibit adaptivity and fault tolerance in the future. The cell signalling and the system of Gene Regulatory Networks present in biological organisms are modelled in our developmental system, and tailored for tackling real world problems on electronic hardware. For the first time, a Gene Regulatory Network system is successfully shown to develop the complete circuit structure of a desired digital circuit without the help of another mechanism or any problem specific structuring. Experiments are presented that show the modular behaviour of the developmental system, as well as its ability to solve non-modular circuit problems.

An Investigation of the Importance of Mechanisms and Parameters in a Multicellular Developmental System

“…As stated earlier, the ADS used in this paper is an already existing model that was designed to enable scalable evolution of digital systems such as circuits [9], [37], controllers [74], and image compression tools [75]. The model that will be described in this section is the original design that was introduced in [9] without any current optimizations or simplifications.…”

“…The model that will be described in this section is the original design that was introduced in [9] without any current optimizations or simplifications. Therefore, the system described will have all the assumptions made in the initial design.…”

“…5 Although more biologically inspired, a micro-model developmental system does not necessarily model biological development perfectly. In fact there is much work in this type of artificial development with diverse design constraints, those that model biology closely [1], [13], [54]- [56], those that aim to model biological development in a simplistic fashion [19], [57], [58], and models that are "in-between" [3], [4], [9], [31], [59]. Mimicking biology closely should provide a developmental system with high evolvability [30], [60], whereas a simplistic model would reduce the number of complicated processes that exist in biological development, reducing simulation times drastically.…”

“…Section III provides a review of some of the relevant ADSs in literature, and provides discussions on some of the various decisions different researchers have taken while designing their systems. Sections IV introduces the ADS used in this paper as it was designed the first time it was introduced in [9]. Section V details the evolutionary algorithm used to optimize the ADS in the experiments presented in this paper.…”

“…The ADS used in this paper is an already existing model that uses a cell chemistry approach to implement development [9]. Hence, the conclusions on the effects of the investigated parameters and mechanisms apply particularly to similar models of development that implement multicellular development at the cell-interactions level; this will be explained further in Section II as "micro-model," which provides a brief background on development.…”