How to Use Project Control 1.0

How to Use

Project Control 1.0

http://odd.enterprises/


Testing First?

12/09/2017 ©odd.enterprises 2

Answering a question before its asked is like creating a solution before a test.



Project Control Model

Solution, Create Test, Create Solution and Test Solution replace traditional control theory boxes.

• Solution replaces Plant

• Test Solution replaces Sensor





Feedforward and Feedback paths are replaced with Create Test and Solution.

• Create Solution replaces Feedback Controller

• Create Test replaces Feedforward Controller





Project Control is based on Control Theory.

• Input, Error, Output and Feedback are used

• Boolean algebra signals are used




PC Inputs and Outputs

Input to model is an obstacle to be solved and output is Solution.

• Solution has testing applied to create feedback

• Difference between input and output is the error




PC Boolean Logic 1

Boolean algebra drives the model.

• Each obstacle and solution are assigned 2 bits

• 1st bit indicates whether a test has been created

• 2nd bit is for a solution




PC Boolean Logic 2

Boolean logic provides indication of progress.

• 00 for no test or solution

• 01 for a solution with no test

• 10 for a test with no solution

• 11 for a test with solution




Input to Model = 11

Identified obstacles are input to the model as 2 bits.

• Input for each Obstacle is 11

• Extend as necessary for any number of Obstacles




Feedback Path = 00

Feedback path indicates the status of a test and solution.

• No test or solution created:

𝐹𝑒𝑒𝑑𝑏𝑎𝑐𝑘 = 00




Input – Feedback = 11 – 00 = 11

Feedback is taken from the Input using bitwise subtraction.

• Actions are determined by Error; as with control theory

• Difference between Input and Output determines actions




Error = 11

Error drives creation of Tests and Solutions.

• 0 / 1 for created / not created

• 1st bit indicates if we need to create a solution

• 2nd bit indicates if we need to create a solution




Error Triggers Test and Solution Creation

Error = 11; therefore we must create a Test and Solution.

• We follow the Feedforwards path first and create test

• We do not create a Solution yet




Test is Created First

Test created first to ensure we know the objective of a Solution.

• Creating a test first helps influence creation of Solution

• Running test with no solution ensures test can fail




Run Test to Ensure it Fails

It is possible that a test can be incorrect and always return a pass.

• Watching a test fail ensures it can fail

• If a test returns a pass without a solution then it is incorrect





Once a test can fail then feedback loop signal becomes 10.

• We now repeat and find the difference between Input and Output.

• Error signal is now equal to 01.




Error = 01

Error = 01; we have a Test and can now create a Solution.

• Create Solution replaces feedback controller

• Solution is designed according to Test




Error Triggers Solution Creation

Error = 01; Test has been created and signals Create Solution.

• We have created a Test and know objective of Solution

• Solution designed according to Test




Solution is Created

Once Solution is created; it is Tested to create Feedback path.

• Solution is tested at earliest possible opportunity

• Once Solution is tested feedback path is complete




Solution Passes Test

Solution passes Test and the Feedback loop is complete.

• Solution passes Test and output from Test Solution = 11

• If Solution doesn’t pass Test then output = 10





Feedback path is complete and we can find Error signal.

• Bitwise subtraction gives Error = 00

• Fully tested Solution has been created




Error = 00

We have created a Test and Solution and

• Error = 00

• Error signal no longer drives the model

• No further need to create Test or Solution




Solution is Tested

Solution has been created according to a test and tested.

• No further development necessary




No Further Development

Error signal = 00 and indicates now further development is necessary.

• Bitwise subtraction of Feedback cancels out Input

• No input to Create Test or Create Solution




Solution Created According to Test

Solution created according to solve the obstacle and we can move to next stage.

• Creating tests provide Verification

• Creating Solution to pas tests provides Validation




Repetition for Each Obstacle

We repeat the same method for each obstacle required to be solved.




Number of Obstacles = 𝑛

Number of obstacles to be solved is denoted by 𝑛.

• Bit length of signals are extended

• Number of bits used = 𝑛 ∗ 𝑑

• Each signal in the model will be equal to 𝑛 bits




Substitute for Number of Obstacles

If you have 4 obstacles then 𝑛 = 4 and 𝑛𝑑 =8 bits.

• 4 bits for the tests

• 4 bits for the solutions

• You can create all the tests first

• You can proceed test then solution, etc




Repeat Previous Steps for 𝑛 Obstacles

How you proceed is up to you.

• You can create all the tests first

OR

• You can proceed test then solution, and repeat

OR

• Use a combination of steps




Repeat until Obstacles are Solved

We repeat the method until all obstacles have a test and solution.

• When Error signal becomes all 0’s we stop development

• Same method is used for all stages, levels and system divisions




Models Exponential Complexity

Other engineering methods become less effective with time and complexity.

• Complexity increases of creating solutions increases exponentially

𝐶𝑜𝑚𝑝𝑙𝑒𝑥𝑖𝑡𝑦 = 2𝑛




Repeat for each Level and Stage

Each stage is created using the same principle of a test and solution.

• Early stages will simply test theoretical ideas

• Early stages ensure we have identified what to create




Full Method for Project Control

We create Obstacles for the next stages and integrate or decompose as necessary.

• When we finish we have made every attempt to achieve success through identifying failure




Further Information and Questions

Website

Presentations

Facebook

Twitter

Email





http://www.odd.enterprises/odd-presentations.html

https://www.facebook.com/odd.enterprises

https://twitter.com/odd_enterprises

mailto:[email protected]

http://www.odd.enterprises/home.html

http://www.odd.enterprises/home.html



Legal Stuff

ReferencesTest Driven Development for Embedded C

James Grenning, 2011

Test Driven Development

http://en.wikipedia.org/wiki/Test-driven development

Behaviour Driven Development

http://en.wikipedia.org/wiki/Behavior-driven development

Unit Testing

http://en.wikipedia.org/wiki/Unit testing

DisclaimerThe ODD M-model and associated processes are provided by odd.enterprises and may be used for any purpose whatsoever.

The names odd.enterprises and associated logos should not be used in any representation, advertising, publicity or other manner whatsoever to endorse or promote any entity that adopts or uses the model and/or associated processes.

odd.enterprises does not guarantee to provide support, consulting, training or assistance of any kind with regards to the use of the model and/or processes including any updates.

You agree to indemnify odd.enterprises and its affiliates, officers, agents and employees against any claim or demand including reasonable solicitors fees, related to your use, reliance or adoption of the model and/or processes for any purpose whatsoever.

The model is provided by odd.enterprises “as is” and any express or implied warranties, included but not limited to the implied warranties of merchantability and fitness for a particular purpose are expressly disclaimed.

In no event shall odd.enterprises be liable for any damages whatsoever, including but not limited to claims associated with the loss of data or profits, which may result from any action in contract, negligence or other tortious claim that arises out of or in connection with the use or performance of the model.




http://en.wikipedia.org/wiki/Test-driven_development

http://en.wikipedia.org/wiki/Behavior-driven_development

http://en.wikipedia.org/wiki/Unit_testing

How to Use Project Control 1.0

Engineering