Models Flashcards

Question

What are the possible types of relations?

Answer 1

Undirected ▪ To express connectedness ▪ For example: spatial, communication Directed ▪ To express dependency, hierarchy, order, … ▪ For example − Necessary / sufficient / INUS condition − Contains (for decomposition) − Flow of material / energy / signals − Temporal order ▪ Dependency is an ambiguous term.

Answer 2

▪Elements such as components, roles, stakeholders, process steps, … ▪Relations between the elements related, e.g., to dependency, hierarchy, connectedness, … ▪Optional: a graph

Answer 3

▪Matrix with elements and relations →input for further analysis (clustering, sequencing) ▪Visualization of system connectivity including type (and optionally intensity) of interaction

Answer 4

▪Presence of manyinteracting or connected elements, related to a product, organization, or process ▪Preparation for further analysis steps: clustering / sequencing

Answer 5

1. Identify and group relevant elements, possibly by decomposing the system. Use as row and column labels. 2. Document the relations, possibly add weighting factor. You can skip the first two steps if you have already a graph as an input and start here: 3. Analyze: Identify patterns and their implications for system behavior (e.g. related to density, distribution, …). 4. Improve: Most DSM applications result not only in a better understanding of the system, but also in the improvement of the system through further analysis (see: clustering or sequencing)

Answer 6

Undirected DSMs (static) and Directed DSMs (temporal)

Answer 7

A DMM is an inter-domain matrix which represents the link between elements from two different domains.

Answer 8

▪Elements from different domains (e.g. components functions, persons-components, persons-topics) ▪Relations between the elements ▪Optional: a graph

Answer 9

▪DMM = tool for mapping in DSM analysis

Answer 10

▪Presence of manyinteracting or connected elements from two different domains, e.g., product-related and actor-related ▪Preparation for further analysis steps: clustering / sequencing

Answer 11

1. Identify and group relevant elements, possibly by decomposing the system. Use elements from domain 1 and 2 as row and column labels. 2. Identify: Document the relations, possibly add weighting factor.You can skip the first two steps if you already have a graph as an input and start here: 3. Apply to further analysis: -multiply with DSMs -extract DSMs from it

Answer 12

▪ The objective of clustering analysis is to organize many elements with pair-wise relationships/dependencies into groups with a small number of relationships/dependencies between separate groups. ▪ This can be done by moving rows and the associated columns simultaneously. ▪ Clustering may be used for modularization of complex products →product family design. ▪ Clustering may be used for grouping people. ▪ DSMs with high connectivity often require algorithmic support.

Answer 13

▪ Unstructured, static DSM ▪ Clustering goal (e.g. numberof clusters, size of clusters, or constraint on relations outside a cluster)

Answer 14

▪ Clusters

Answer 15

▪ Presence of overwhelmingly many elements with undirected relations (e.g. components or team members)

Answer 16

1.Get an overviewand build intuition. 2. Isolate extreme elements, i.e., elements that have: -connections to almost every other element (highly connected elements) -almost no connection to other elements by moving them into the upper left or lower right corner and leave them out of further consideration. 3.Clusterremaining elements into chunks: find elements which have similar entries and move them together by switching columns and rows simultaneously 4.Markclusters with a rectangle for better visualization. All entries which are not included in the cluster cause interactions between clusters. 5.The aim is to minimize interactions between clusters →Repeatfrom step 3 if necessary

Answer 17

The DSM lists all constituent elements and the corresponding information flow, interactions and dependency patterns of a process architecture. -Input into columns (from rows), feedback below diagonal

Answer 18

▪ Sequencing is the reordering of rows and columns of a DSM with the objective of arranging all existing dependencies on the upper side or at least as close as possible to the diagonal. ▪ Sequencing produces smaller feedback loops → process steps are closer to each other. ▪ A triangular matrix on the top, i.e. all dependencies on the upper side of the diagonal, represents a process without any feedback. ▪ If there are no feedback loops, sequencing will produce a triangular matrix.

Answer 19

▪ Process planning ▪ Presence of processes with −Large feedback loops −Many feedback loops

Answer 20

▪Unstructured, temporal DSM Matrix

Answer 21

▪ Feedback loops ▪ Process with minimized feedback loops.

Answer 22

1.Get an overviewand build intuition. 2. Identify elements which do not depend on information of other elements (no entries in column) and move them to the upper left corner 3. Identify elements which do not influence elements (no entries in rows) and move them to the lower right corner (in Excel this can be done by pressing shiftwhile moving). If there are no remaining elements in DSM after steps 1 and 2, the matrix is completely sequenced, otherwise the remaining elements contain information feedbacks (at least one). 4.Decrease the size of feedback loops by switching elements with entries far away from the diagonal in order to move entries closer to the diagonal.This results in isolated and smaller feedback loops. Note: Process tasks which have entries on both sides of the diagonal are circular dependencies. They cannot be removed, but only isolated and reduced

Answer 23

Grouping principle Functional organization: By expertise, competence, and experience Project organization: By product, product function, or development goal Typical view Functional organization: Deep specialization; discipline-centric perspective Project organization: Integrated product perspective; cross-functional balance Decision speed Functional organization: Slower due to coordination across multiple expert groups Project organization: Faster because decisions are made closer to the product Knowledge reuse Functional organization: High reuse of domain knowledge and standards Project organization: Reuse limited Accountability Functional organization: Distributed across specialist groups Project organization: Clear product responsibility within the team

Answer 24

▪ A matrix organization is a hybrid structure where individuals have organizational links to both, projects and (organizational) functions at the same time. →Combines strengths of both, functional and project view ▪ Individuals typically have several supervisors. →Requires more management and administration

Answer 25

-Product (function) designer: Responsible for customer-relevant attributes of product -Component designer: Responsible for component attributes -Architect: Responsible for system attributes, essential for product architecture -Project manager: Responsible for meeting the project goals

Answer 26

Product architecture is a scheme by which functional elements of the product are arranged into physical chunks and by which the chunks interact with each other

Answer 27

▪ Consists of elements (nodes) that represent attributes (= properties), here A, B, C. Ideally, attributes are objectively measurable, i.e., two different observers obtain the same value. ▪ Has relations (edges) that represent dependency in a particular sense: − Values for C can only exist, when values for A & B exist: attributes A & B are measurable before attribute C. − Values of attributes A & B determine the value of C by the laws of physics (rather than by some decision maker) →The value of C is a physical consequence of the values of A & B. − Example: A and B are attributes of separate components 1 & 2. Only after joining components 1 & 2, the attribute of system attribute C emerges, i.e., C can be measured. ▪ Should contain all attributes that (1) may change value during design or use of the product and (2) are helpful to formulate INUS conditions → building plan for requirement derivation. ▪ Must not contain requirements! ▪ Which attributes are treated as QoIs and DVs depends on the scope.

Answer 28

-direct influence when there is no other attribute in between -indirect influence when there is an intermediate attribute

Answer 29

▪ Ambiguous dependencies: between quantities that are related by an equation that does not express a requirement, without obvious / relevant order of occurrence → choose quantity of interest according to practical aspects ▪ Dependencies between potentials / capacities

Answer 30

Quality Function Deployment (QFD) has a two-level view − Easier to create, visualize and communicate − May or may not include all relevant information Attribute Dependency Graph (ADG) has a multiple level view − Resolves dependencies in detail − More modelling effort

Answer 31

▪Scope of analysis expressed by (1) requirements and/or quantities of interests and/or (2) design variables involved. ▪Rough qualitative understanding of dependencies

Answer 32

▪Dependency structure of attributes ▪Building plan for requirement development ▪Candidates of causes of a problem

Answer 33

▪Presence of many attributes and dependencies ▪Need to structure a problem and decompose it into sub-problems ▪Organize interfaces between several disciplines ▪Preparation for requirement formulation

Answer 34

1. Forget all requirements, needs, intentions and purpose (however, include the associated attributes) 2. List all attributes that may be of relevance (QoIs, DVs, intermediate attributes) 3. Order attributes by dependency = order of emergence of properties. Assume view to assemble or build a product by choosing arbitrary values (to ensure that you do not use requirements) and observe how properties emerge by assigning values to attributes. 4. Check direct influence: Is it possible to change C by changing B without changing A? If YES: there is a direct influence. 5. Do you improve clarity by adding intermediate attributes?

Answer 35

Risk of silo thinking and sub-optimization within specialties

Answer 36

Risk of inconsistent methods or standards across products

Models Flashcards

(60 cards)