CDS Data Analysis Inc. - Context Integrated System Optimization -

Home C.I. Data Analysis C.I. System Optimization Contact / Comment

Context Integrated System Optimization

Few systems used to model real life production systems take into account all the interconnectivity and interferences that actually occur.
Though they might be apparent and quantifiable, the complexities of systems that include the entire context created by these features quickly overwhelm the capacity of most available optimization systems.
Reduction and filtering are therefore unavoidable processes that essentially enable standard optimization processes.
While many occurrences of quantifiable features might seem negligible under most circumstances, systems regularly collapse exactly because of the employed filtering and reduction processes. Data that can indicate mounting stress early on is often available in the pre-filtering scenario and falls victim to the prerequisites of applied optimization processes.

In many production processes, the interactions between different aspects are well known and understood. However, due to the complexity of the interactions, it was often not possible to optimize such systems.

This is why CDS Data Analysis Inc. developed a new approach that acquires insight in systems where interaction between different domains is completely unknown and only defined domains and observed data thereof are available. Based on observed changes of the systems due to outside influences, strategies are developed to use these influences to improve the overall situation or efficiency of the entire system.

The graphic to the left shows the control unit of a highly complex system. It is comprised of 15 domains of which only eight can directly be influenced. All domains are influential on each other. These influences are defined through 36 non-linear functions. The circle diagram shows the overall situation of the system at the start and at the end of the optimization process. Each sector represents one domain. Fully coloured sectors represent the best possible state the respective domain can be in.

Note: Due to the influence the various domains have on each other, it is not possible to improve all domains to the maximum. To achieve the improvement of the overall situation as shown in the graphic above, a strategy comprising 37 steps, including 7 setbacks (temporary worsening of the overall situation) had to be developed. The developed strategies are often non-trivial. Temporary worsening of the overall situation has to be accepted in order to achieve the ultimate overall improvement.

The graphic below shows the domains of the complex system above and the connections between the domains. Each arrow represents a non-linear function describing the actual influence. Three arrows do not end at a domain. These functions serve to amplify the effect of the influence described by the arrow they point to.


Home C.I. Data Analysis C.I. System Optimization Contact / Comment
Context Integrated System Optimization Few systems used to model real life production systems take into account all the interconnectivity and interferences that actually occur. Though they might be apparent and quantifiable, the complexities of systems that include the entire context created by these features quickly overwhelm the capacity of most available optimization systems. Reduction and filtering are therefore unavoidable processes that essentially enable standard optimization processes. While many occurrences of quantifiable features might seem negligible under most circumstances, systems regularly collapse exactly because of the employed filtering and reduction processes. Data that can indicate mounting stress early on is often available in the pre-filtering scenario and falls victim to the prerequisites of applied optimization processes. In many production processes, the interactions between different aspects are well known and understood. However, due to the complexity of the interactions, it was often not possible to optimize such systems. This is why CDS Data Analysis Inc. developed a new approach that acquires insight in systems where interaction between different domains is completely unknown and only defined domains and observed data thereof are available. Based on observed changes of the systems due to outside influences, strategies are developed to use these influences to improve the overall situation or efficiency of the entire system.
	The graphic to the left shows the control unit of a highly complex system. It is comprised of 15 domains of which only eight can directly be influenced. All domains are influential on each other. These influences are defined through 36 non-linear functions. The circle diagram shows the overall situation of the system at the start and at the end of the optimization process. Each sector represents one domain. Fully coloured sectors represent the best possible state the respective domain can be in.
Note: Due to the influence the various domains have on each other, it is not possible to improve all domains to the maximum. To achieve the improvement of the overall situation as shown in the graphic above, a strategy comprising 37 steps, including 7 setbacks (temporary worsening of the overall situation) had to be developed. The developed strategies are often non-trivial. Temporary worsening of the overall situation has to be accepted in order to achieve the ultimate overall improvement.
The graphic below shows the domains of the complex system above and the connections between the domains. Each arrow represents a non-linear function describing the actual influence. Three arrows do not end at a domain. These functions serve to amplify the effect of the influence described by the arrow they point to.