Week 13 - Process Redesign (3) and Implementation

Source / Reference:
1)"Pattern-based reasoning for rapid redesign: a proactive approach " by Li, Simon ; Chen, Li, Research in Engineering Design, 2010, Vol.21(1), pp.25-42
http://www.springerlink.com.ezproxy.lb.polyu.edu.hk/content/k523937104546118/


Subject:
Further development of Rapid Redesign model: a proactive approach


Motivation:

Refer to my previous blog: Other model of process redesign: Model-based Rapid Redesign, I further investigate more on his topic and I found another article that is written by the same author, which is a further development of the rapid design model. In the successive article, a proactive approach, Pattern-based reasoning for rapid redesign is discussed. This blog is to briefly introduce this new approach, compare it with the previous one and observe what have been improved.


Introduction
In the original approach - Pattern-based reasoning for rapid redesign, the decomposition process for redesign is not activated until the presence of a redesign request. This prior work represents a reactive approach where a new set of decomposition patterns should be generated in accordance with a different redesign request input.
In the new proactive approach, the decomposition patterns capturing generic decomposed structures of a given design model are created in advance and stored in a design library before any redesign request emerges. These pre-generated patterns are able to address any upcoming redesign request without further decomposition procedures in redesign. This proactive approach is developed in a new framework of pattern based reasoning that is built on the mechanism “case  pattern  strategy”. Two methodological components, Proactive Redesign Decomposition and Redesign Condition Analysis, are introduced in the article.


Proactive approach
The pattern-based reasoning mechanism discussed in previous blog is treated as a knowledge base to support the development of the rapid redesign methodology. Particularly, redesign cases and scenarios support the formation of pattern solutions for any redesign problems represented by DDM. Then, redesign strategies provide a finite range of tactics that help to systematically develop a detailed redesign roadmap for any redesign scenario. There are two components for the proactive redesign approach: Proactive Redesign Decomposition and Redesign Condition Analysis.

Proactive redesign decomposition
The development of Proactive Redesign Decomposition is based on the two-phase decomposition method. The first phase of this method, termed dependency analysis, is applied such that the scattered ‘‘1’’ elements in the original DDM get close to one another to form clusters. For most of complex problems in practice, the banded diagonal matrix will be obtained, which indicates the existence of interactions between the formed clusters. After obtaining the banded diagonal matrix, the second phase, termed matrix partitioning, is invoked. In this phase, three decomposition criteria are required in order to generate decomposition solutions: number of blocks, size limit of each block, and size limit of interaction parts.
The pattern complexity metric is a relative measure that estimates the potential design (computing) effort associated with a model-specific pattern. The block size balance is measured via the SD of the block sizes. Both of these values are used to analyze and refine the generated set of decomposition solutions, thus finally forming the model-specific pattern library


Redesign condition analysis
After obtaining the model-specific pattern library, the proactive redesign approach is ready to receive redesign requests, which activate the Redesign Condition Analysis. In this analysis, the incoming redesign requests are interpreted and translated into the corresponding target entities in the DDM. Then, the target entities are labeled in the model-specific patterns to form pattern solutions. Also, it is required to identify the redesign cases in order to select the proper root patterns from the model-specific pattern library. As a result, the applicable pattern solutions are identified to form the pattern selection space. At this point, the Redesign Planning Analysis can be applied to finally generate the redesign roadmap for the guidance of the redesign solution process.


Conclusion
The proactive redesign approach in this paper utilizes the two-phase decomposition method to generate decomposed structures of a given model to approximate its potential pattern solutions. Thus, as long as the existing model is kept the same, no further decomposition is required for new redesign requests. This proactive technique is applicable to existing design models that are subject to frequent yet minor design changes.
In short, the proactive redesign can generate a library that contains the best combination of process patterns that allowing certain changes. If we find some process sequence needed to be change, other patterns can be used to replace the existing method instantly since all the calculation is done preliminary. Compare with the previous approach, the proactive redesign can actually pay some efforts before the redesign is conducted. Also, once the computation is done, it can be reused in the future subject to no rapid change to the process dependency and sequence.