Project

Silicon Lifecycle

Goal: Within the next 5 years, we envisage cheap silicon attached to most manufactured items (even inert things like steel plates) that:
• Identifies itself (subsuming RFID),
• Geo-locates itself using GPS and Wireless triangulation
• Has substantial memory and processor power (substantial defined
later),
• Is wireless, using several modes including ultra-wideband,
• Incorporates sensors (via MEMs on the chip, externally via plug ‘n
play sensors e.g. IEEE 1451, or embedded in a larger smart
material matrix, and
• Is attached for the total lifecycle of an asset, from its creation to its
destruction.

We call this the Silicon Lifecycle. This critical mass of cheap, small silicon will cause “Innervation”, defined here as the act of innerving or stimulating as with the distribution of nerves in an animal, to and from its brain, spinal cord and all of its parts. Substantial computing with wireless bandwidth to and from every critical asset will bring the business intelligence to the last mile for most every manufacturing endeavor. It will incorporate:
– self-healing, self-organizing, peer-to-peer, web-services based grid
computing amongst networks of connected assets, optimization
of actions using, as a basis, reinforcement learning (a form of
dynamic programming)
– business logic execution
– real options for decision valuation via dynamic programming
supplied by reinforcement learning memory so it can capture its
own best-practices and be its own historian

In summary, innvervation will be enabled by adaptive, anticipatory
control based reinforcement learning and evolutionary algorithms.

Updates
0 new
1
Recommendations
0 new
0
Followers
0 new
5
Reads
0 new
23

Project log

Roger N. Anderson
added a research item
Balancing environmental, economic, and societal needs for a sustainable future encompasses problems of unprecedented size and complexity. Computing and information science can — and should — play a key role in addressing critical sustainability challenges faced by present and future generations.
Albert Boulanger
added a research item
Conversion to lean energy management will be required to insure economic development in ultradeep-water regions. Software-controlled lean systems integration will drive innovation toward breakthrough cost and cycle-time savings. This move will utilize commercial off the shelf (COTS) software widely used in the communications, chemical processing, aerospace, automotive , and other high-tech manufacturing industries, but not yet used for lean management in the upstream energy industry.
Albert Boulanger
added an update
See this research direction paper.
And this presentation.
 
Albert Boulanger
added 5 research items
The purpose of Computer Aided Lean Management (CALM) is to enable operational innovation through the deployment of software algorithms. Lean management is a methodology for efficient enforcement of process rigor and discipline in order to dramatically cut costs and improve operations of an enterprise (see http://leanenergy.ldeo.columbia.edu/ogj). This software development will also reduce operating risk, enhance customer service and reliability, and increase the assurance that a new design introduced to the "market" will be effective. CALM is software-controlled lean systems integration that drives breakthroughs in cost and risk reduction. Operational innovation within an energy organization will be enabled through the integrated deployment of three major software systems that we call the Integrated System Model (ISM):  Product modeling-High resolution model of physical infrastructure.  Business process modeling-Capturing detailed process and work flow information in order to track and measure performance on a daily basis with a goal of optimizing these processes.  Machine learning system-Diagnostic analysis of historical and operational data captured in existing data as well as from Product model and Business Process model outputs to predict and/or prioritize required operations and maintenance of an energy company's business. CALM is a methodology for running a business based on the common sense approach of measuring the results of actions taken and using those measurements in an experimental way to design new processes that drive out inefficiencies. In the ISM we will have models of the business where alternatives can be explored to find the innovations required to improve the company's performance. The ISM will provide the tools needed to "see" the competitive landscape or environment the company operates in. Some the feedback to improve performance will be provided by the machine learning tools being developed in this project. The company will need these tools and will need to adopt CALM in order to become more adaptive and therefore better able to perform successfully in the future as the "business we are in" changes.
A Calculus of Value based on Ubiquitous Silicon Inspired by Nature We foresee every value-based decision to use this calculus of value under uncertainty -- even a pump can know its real options and its synergies (covariance) with other devices. • The silicon will be there and cheap with ample compute power to implement this at the smallest levels. • Real Options well be extended to Dynamic Real Options – i.e. the silicon will dream up new options. – The dream phase of the adaptive control algorithms uses simulation and genetic algorithms to come up with new options. – Deep ties with Quantum Mechanics and variational principles in physics • This may sound like fantasy, but it is all very realizable and a large body of work already developed in separate areas. Need to put it all together
Albert Boulanger
added a project goal
Within the next 5 years, we envisage cheap silicon attached to most manufactured items (even inert things like steel plates) that:
• Identifies itself (subsuming RFID),
• Geo-locates itself using GPS and Wireless triangulation
• Has substantial memory and processor power (substantial defined
later),
• Is wireless, using several modes including ultra-wideband,
• Incorporates sensors (via MEMs on the chip, externally via plug ‘n
play sensors e.g. IEEE 1451, or embedded in a larger smart
material matrix, and
• Is attached for the total lifecycle of an asset, from its creation to its
destruction.
We call this the Silicon Lifecycle. This critical mass of cheap, small silicon will cause “Innervation”, defined here as the act of innerving or stimulating as with the distribution of nerves in an animal, to and from its brain, spinal cord and all of its parts. Substantial computing with wireless bandwidth to and from every critical asset will bring the business intelligence to the last mile for most every manufacturing endeavor. It will incorporate:
– self-healing, self-organizing, peer-to-peer, web-services based grid
computing amongst networks of connected assets, optimization
of actions using, as a basis, reinforcement learning (a form of
dynamic programming)
– business logic execution
– real options for decision valuation via dynamic programming
supplied by reinforcement learning memory so it can capture its
own best-practices and be its own historian
In summary, innvervation will be enabled by adaptive, anticipatory
control based reinforcement learning and evolutionary algorithms.