Distributed Hardware Evolution/en
 |
Languages | • • |
Self-diagnostic circuits have the ability to monitor themselves and find deviations from their normal behavior. Such skills are particularly needed for very critical and dangerous ventures, for example in space or in the control of power plants.
Generally, to realize such capabilities, the actual circuit is constructed in duplicate (or even with even greater redundancy). The results of both circuits are compared to each other and, in case of deviation, it is decided with the help of a poll which result is used. This approach means that more than twice the chip area is needed to make the circuit rather than the simple design. The research in the field of such self-diagnostic systems has, in addition to these well-known multi-redundant systems, so far hardly came up with new techniques what should be changed by this project.
Simple circuits are further developed using a genetic algorithm. For this purpose, all properties of the circuit are binary coded. For starters, random strings representing random circuits are chosen. For all these circuits it is now determined how well they can solve the problem posed (simple arithmetic operations or the like). The best circuits are now selected and "paired", i.e. new circuit codes are determined, which are put together from the best current circuits.
In order for this development to not stall too fast, some circuits, i.e. in their encodings, one or more bits randomly flip over. In this way, over time, circuits are created that are always better adapted to the problem.
The project run from 2003 till 2005 with the same name on another server: Archive
DHEP was reactivated in May 2018 by the University of Sussex (UK).
Badges are beeing designed. [1]
Inhalt
Project overview
 Distributed Hardware Evolution
| |
|---|---|
| Name | Distributed Hardware Evolution |
| Category | Physik/en |
| Goal | Development of self-diagnostic circuits |
| Commercial | no |
| Homepage | dhep.ga/boinc/ |
Project status
| This project is being conducted in Great Britain. |
|
Department of Informatics University of Sussex, Großbritannien |
Project links
- About the project
- Create an account
- Server status
- Statistics
- Forum
- Project scientist: Dr. Michael Garvie
- Scientific website by Dr. Michael Garvie
Statistics
| Where | Overview | Top Teams | Top User |
|---|---|---|---|
| Project Home Page | Statistics goal | Top Teams | Top User |
Client properties
| works also with proxy | 
|
| program can be executed normally | 
|
| can be used as screensaver |
|
| commandline version available |
|
| personal Proxy for work units available |
|
| work units can be exchanged also by email |
|
| source code is available |
|
| can be used offline |
|
| checkpoints |
|
The information above is not fully checked.
Checkpoints: Every 15 minutes a so-called trickle-up information is sent to the server. Data and credits are therefore reported every 15 minutes.
The WUs are dummy placeholders in BOINC, i.e. neither the total time remaining nor the progress has any significance. Just run and use your pc as desired.
News
RSS-Feed
Der RSS-Feed von https://dhep.ga/boinc/rss_main.php konnte nicht geladen werden: Fehler beim Abruf der URL: SSL: no alternative certificate subject name matches target host name 'dhep.ga'
Selected News
- News Scientist, 31.7.2003: "Circuit design evolving in distributed digital world"
- Fachbeitrag, Oktober 1996: "An evolved circuit, intrinsic in silicon, entwined with physics." (PDF file)
Quality control
Checked: 20.10.2018
