Network Analysis and deeper understanding of complex systems: PositivesNetwork Analysis and deeper understanding of complex systems II: NegativesSchematic of Engineering System Model PurposesNetwork Model types of interest (from lecture 12)Network Analysis and deeper understanding of complex systems III: A possible future approachComparative Progress in Understanding and performance: CLM objective/subjective observationsNetwork Analysis and deeper understanding of complex systems IIThe Iterative Learning ProcessComparative Progress in Understanding and performance: CLM objective/subjective observationsNetwork Analysis and deeper understanding of complex systems IIProfessor C. Magee, 2006Page 1Network Analysis and deeper understanding of complex systems: Positives• Some helpful general insights: For example “Small Worlds are Ubiquitous” in large (lots of nodes or links) systems• Network Analysis is a methodology that nicely crosses over between technical concerns (much sociology work) and technical matters (OR and other extensive network approaches). This crossover or dualism is at the root of Engineering Systems issues as we need to develop quantitative methods which treat both aspects.• Comparative studies of structures in very different systems and the development of “observation” techniques to arrive at some potentially important structural information from a network model.• Community analysis• Motif analysis• Centrality and other social structure metrics• Some superb models have been developed for search, navigation, organization design and distribution systems. A characteristic of these exemplar modeling efforts has been that they all combine previously separate sociology, and OR/CS approaches.Professor C. Magee, 2006Page 2Network Analysis and deeper understanding of complex systems II: Negatives• Overgeneralization (as happened 50 years ago in “General Systems Theory”). The clearest example was (is?) the tendency to assume that power laws of degree distribution imply “Scale-Free” Structures because one model that is consistent with power laws leads to such structures. However, it is clear now (and known much earlier) that multiple models lead to power laws. It is also now clearer that power laws are “More Normal than Normal” distributions when variation is not bound.• The + r, - r story for sociological vs technological systems also appears to be premature generalization but this has not resulted in the “cottage industry” that power laws did nor the mental block that “scale-free” has raised.• Not nearly enough attention to data quality and facts: “it isn’t what you don’t know that gets you in trouble but what you think you know that just ain’t so that does.”• Methods for handling really large datasets are needed• Not enough attention to realistic “System Formation Models”Professor C. Magee, 2006Page 3Schematic of Engineering System Model PurposesSystem StructureQuantified by aRich set of metricsSystem Propertiesunderstood quantitatively in terms of desirabilitySystemformation mechanisms andconstraintsNetwork math models to predictproperties from structure(System processes andproperties models)Network math models to predict structure (System formation models) Architecture representedas networks (Systemobservation models)Professor C. Magee, 2006Page 4Network Model types of interest (from lecture 12)• Models/algorithms used to “observe” systems • Calculation of structural metrics• Communities, motifs, coarse-graining, hierarchy• Models for predicting/explaining Structure• Models for formation/growth processes of systems• Most network models such as random, small-world etc. implicitly fall in this category• Cumulative advantage, preferential attachment, bipartite community formation, heuristic optimization relative to constraints, hierarchy (or heuristics) + random• Models for predicting/explaining properties of systems• Predicting properties from structure – architecture • Flexibility, robustness, performance of functions • Operational processes or functions• Communication, problem solving, decision-making, learning• Search and navigation• Failures and cascades, epidemicsLecture 6,7, 8 and 18Lectures 7,10,12, 14, 15, 18 and 20Lectures12, 14, 17, 18 and 20SociologyORSociology,EngineeringCS & ORProfessor C. Magee, 2006Page 5Network Analysis and deeper understanding of complex systems III: A possible future approach• Utilization of the framework espoused here should help keep the areas of ignorance and the need for models of use in design more clearly in focus.• Why should we care?• How much progress have we made in complex system design (for example organizational design) from 1940-2000?Professor C. Magee, 2006Page 6Comparative Progress in Understanding and performance: CLM objective/subjective observations• 1940-2000 improvement• Small-scale electro-mechanical systems (x40-100)• Energy transformation systems (x 10-20)• Information processing systems (x to )• Cosmology (x 30-100)• Paleontology (x 50)• Organizational theory and practice (x 1.1 to 2)• Economic systems (x 1.1 to 2)• Complex large-scale socio-technological systems (?)12101510Professor C. Magee, 2006Page 7Network Analysis and deeper understanding of complex systems II• Why should we care?• How much progress have we made in complex system design (for example organizational design) from 1940-2000?• What is needed to greatly improve the practice of complex social/ technological system design?• The major opportunity is to transition from the “pre-engineering” (experiential or craft) approach now widely used to a solid (post 1870 at least) engineering approach to these design problems. This basically involves moving our learning in these areas so that a “Cumulative” Science evolves (rather than “paradigm of the month science). What does this entail?Professor C. Magee, 2006Page 8The Iterative Learning Processdeduction induction deduction inductionObjectively obtained quantitative data (facts, phenomena)hypothesis ( model, theory that can be disproved)As this process matures, what new can the models accomplish?The major accomplishment will be the rapid facilitationof a transition to engineering (vs. craft approaches) for thedesign of complex social/ technological systemsProfessor C. Magee, 2006Page 9Comparative Progress in Understanding and performance: CLM
View Full Document
Unlocking...