Fast Hybrid Genetic Clustering Algorithm for Design Structure Matrix

2013 ◽  
Author(s):  
Fredrik Borjesson ◽  
Ulf Sellgren
Keyword(s):  
Clustering Algorithm ◽  
Minimum Description Length ◽  
Clustering Algorithms ◽  
Optimal Solution ◽  
Design Structure Matrix ◽  
Computer Algorithms ◽  
Structure Matrix ◽  
Design Structure ◽  
Genetic Clustering ◽  
Time Required

Module formation is the step in which a product’s architecture is established in such a way that complex interactions are intra-modular and inter-modular interactions are more simple. If a matrix representation exists, such as a Design Structure Matrix, this involves clustering system entities into groups with strong intra-dependencies. For simple products, clustering may be done manually, but for complex products, computer tools are required. Existing clustering algorithms are either slow, or unable to guarantee a globally optimal solution. To enable iterative work and to make cluster analysis useful also in the detailing steps, efficient and effective computer algorithms are required. This paper presents an efficient and effective Genetic clustering algorithm, with the Minimum Description Length measure. To significantly reduce the time required for the algorithm to find a good clustering result, a knowledge aware heuristic element is included in the GA process. The efficiency and effectiveness of the algorithm is verified with four case studies.

Improved Clustering Algorithm for Design Structure Matrix

2012 ◽  
Author(s):  
Fredrik Borjesson ◽  
Katja Hölttä-Otto
Keyword(s):  
Clustering Algorithm ◽  
Hill Climbing ◽  
Design Structure Matrix ◽  
Computationally Efficient ◽  
Vacuum Cleaner ◽  
Local Optima ◽  
Structure Matrix ◽  
Design Structure ◽  
Clustering Problems

For clustering a large Design Structure Matrix (DSM), computerized algorithms are necessary. A common algorithm by Thebeau uses stochastic hill-climbing to avoid local optima. The output of the algorithm is stochastic, and to be certain a very good clustering solution has been obtained, it may be necessary to run the algorithm thousands of times. To make this feasible in practice, the algorithm must be computationally efficient. Two algorithmic improvements are presented. Together they improve the quality of the results obtained and increase speed significantly for normal clustering problems. The proposed new algorithm is applied to a cordless handheld vacuum cleaner.

Incorporating Constraints in System Modularization by Interactive Clustering of Design Structure Matrices

2016 ◽  
Author(s):  
Roozbeh Sanaei ◽  
Kevin N. Otto ◽  
Katja Hölttä-Otto ◽  
Kristin L. Wood
Keyword(s):  
Complex Systems ◽  
Clustering Algorithm ◽  
Clustering Algorithms ◽  
Design Structure Matrix ◽  
Complex Products ◽  
Pressure Fields ◽  
Design Structure ◽  
Computational Capability ◽  
New Type ◽  
Interactive Clustering

Modularity is an approach to manage the design of complex systems by partitioning and assigning elements of a concept to simpler subsystems according to a planned architecture. Functional-flow heuristics suggest possible modules that have been demonstrated in past products, but using them still leaves it to the designer to choose which heuristics make sense in a certain architecture. This constitutes an opportunity for a designer to take other constraints and objectives into account. With large complex systems, the number of alternative groupings of elements into modular chunks becomes exponentially large and some form of automation would be beneficial to accomplish this task. Clustering algorithms using the design structure matrix (DSM) representation search the space of alternative relative positioning of elements and present one ideal outcome ordering which “optimizes” a modularity metric. Beyond the problems of lack of interactive exploration around the optimized result, such approaches also partition the elements in an unconstrained manner. Yet, typical complex products are subject to constraints which invalidate the unconstrained optimization. Such architectural partitioning constraints include those associated with external force fields including electric, magnetic, or pressure fields that constrain some functions to perform or not perform in different regions of the field. There are also supplier constraints where some components cannot be easily provided with others. Overall, it is difficult to simply embed all objectives of modular thinking into one metric to optimize. We develop a new type of interactive clustering algorithm approach considering multiple objectives and partitioning constraints. Partitioning options are offered to a designer interactively as a sequence of clustering choices between elements in the architecture. A designer can incorporate constraints that determine the compatibility or incompatibility of elements by choosing among alternative groupings progressively. Our aim is to combine computational capability of clustering algorithms with the flexibility of manual approaches. Through applying these algorithms to a MRI machine injector, we demonstrate the benefits of interactive cooperation between a designer and modularity algorithms, where constraints can be naturally considered.

scholarly journals An Enhanced Collaborative Optimization Approach with Design Structure Matrix Algorithms to Group and Decouple Multidisciplines

2016 ◽  
Vol 2016 ◽  
pp. 1-8 ◽  
Author(s):  
Xifan Yao ◽  
Jifeng Zhou ◽  
Yongxiang Li ◽  
Erhui Liu
Keyword(s):  
Multidisciplinary Design Optimization ◽  
Clustering Algorithms ◽  
Collaborative Optimization ◽  
Design Structure Matrix ◽  
Optimization Approach ◽  
Structure Matrix ◽  
Matrix Algorithms ◽  
Design Structure ◽  
Solution Efficiency ◽  
Efficiency And Reliability

To improve the solution efficiency and reliability of multidisciplinary design optimization (MDO), an enhanced MDO approach, called sequenced collaborative optimization (SCO), is proposed. The proposed approach introduces the design structure matrix (DSM) to describe the coupling effects among disciplines and aggregates those mutually coupling disciplines into the strong tie groups among similar ones and the weak tie among heterogeneous ones through clustering algorithms. Further, those in the same group are sequenced by the DSM division algorithm. Moreover, by adding constraints, the groups are made independent, resulting in a tree structure without loops, thus decoupling the original multidisciplinary problem into several independent collaborative optimization modules. In the end, an example is employed to verify the efficiency and reliability of the approach.

Multi-Level Flow-Based Markov Clustering for Design Structure Matrices

2016 ◽  
Author(s):  
Tim (T.) Wilschut ◽  
Pascal (L. F. P.) Etman ◽  
Jacobus (J. E.) Rooda ◽  
Ivo (I. J. B. F.) Adan
Keyword(s):  
Clustering Algorithm ◽  
Visual Representation ◽  
System Structure ◽  
Underlying Structure ◽  
Design Structure Matrix ◽  
Structure Matrix ◽  
Design Structure ◽  
Input Parameters ◽  
Markov Clustering ◽  
Multi Level

For decomposition and integration of systems one requires extensive knowledge on system structure. A Design Structure Matrix (DSM) can provide a simple, compact and visual representation of dependencies between system elements. By permuting the rows and columns of a DSM using a clustering algorithm, the underlying structure of a system can be revealed. In this paper, we present a new DSM clustering algorithm based upon Markov clustering. The developed clustering algorithm is able to cope with the presence of ‘bus’ elements, returns multilevel clusters, is capable of clustering both directed as well as undirected DSMs, and allows the user to control the cluster results by tuning only three input parameters.

Analysis of Engineering Change Based on Product Development Network

2011 ◽  
Vol 314-316 ◽  
pp. 1607-1611
Author(s):  
Zhong Wei Gong ◽  
Hai Cheng Yang ◽  
Rong Mo ◽  
Tao Chen
Keyword(s):  
Product Development ◽  
Design Structure Matrix ◽  
Complex Activity ◽  
Engineering Change ◽  
Structure Matrix ◽  
Manufacturing Enterprises ◽  
Design Structure ◽  
Development Network ◽  
Motorcycle Engine

Engineering change is an important and complex activity for manufacturing enterprises. In order to improve the efficiency of engineering change, designers should pay different attentions to different nodes of product development network. In that case, a method of classifying the nodes was proposed. First, we proposed a method to cluster the nodes based on design structure matrix; then, we analyzed the indexes for evaluating the importance of nodes and studied the method of classifying the nodes of product development network; finally, the experiment of managing a type of motorcycle engine was employed to validate our method and it showed the correctness of the proposed method.

Sign in / Sign up

Export Citation Format

Share Document