Heterogeneity and Network Structure in the Dynamics of Diffusion: Comparing Agent-Based and Differential Equation Models

Rahmandad, Hazhir; Sterman, John D.

doi:10.1287/mnsc.1070.0787

Cited by 552 publications

(171 citation statements)

References 70 publications

Supporting

Mentioning

166

Contrasting

Unclassified

Order By: Relevance

“…They note that especially the transformation between the two levels is challenging. Rahmandad & Sterman (2008) (Forrester, 1958). This aligns well with the aims of this work.…”

Section: Resultssupporting

confidence: 88%

See 1 more Smart Citation

Structure‐Based System Dynamics Analysis of Engineering Design Processes

Kasperek

Schenk

Kreimeyer³

et al. 2016

Systems Engineering

View full text Add to dashboard Cite

The dynamic process behavior of Engineering Design Processes (EDPs) needs to be understood in order to distribute resources appropriately as well as for cost and schedule calculation. Currently, structural and behavioral modeling approaches are used in the context of EDP management. In many use cases, it is essential to have at hand both the static structural view and the dynamic behavioral one. Currently, however, the two modeling approaches cannot be sufficiently combined. Within this work, a framework for the structure‐based System Dynamics analysis of EDPs is presented. It uses structural models as a basis for constructing System Dynamics models to simulate the process behavior and thereby supports understanding the correlation of process structure and behavior. This framework was iteratively developed by conducting case studies and consists of three layers, a five‐step procedure and special support for the transformation of the models. For the validation, a subset of the design process at a large German commercial vehicle manufacturer has been chosen. Based on the application of the framework during a six‐month project at the company, the industrial experts rated the framework as beneficial for the purpose of optimizing EDPs. Furthermore, they stated that the transparent representation of the sequence of the process, influencing factors and control variables in combination with the quantitative data had improved their understanding of the overall process.

show abstract

“…They note that especially the transformation between the two levels is challenging. Rahmandad & Sterman (2008) (Forrester, 1958). This aligns well with the aims of this work.…”

Section: Resultssupporting

confidence: 88%

“…A significant variation in output can be caused by parameter uncertainty (Rahmandad & Sterman, 2008). This was especially spotlighted by the scenario analysis in this case study.…”

Section: Exposure To Parameter Uncertaintymentioning

confidence: 77%

Structure‐Based System Dynamics Analysis of Engineering Design Processes

Kasperek

Schenk

Kreimeyer³

et al. 2016

Systems Engineering

View full text Add to dashboard Cite

show abstract

“…(e.g., Urban et al 1990). However, compartment models still assume homogeneity and perfect mixing within compartments and do not consider heterogeneity in individual attributes and in the network structure of interactions (Rahmandad and Sterman 2008). For the Bass model, efforts to explain changes in parameter estimates due to underlying heterogeneity of the population were also made (e.g.…”

Section: Aggregate Models Of Innovation Diffusionmentioning

confidence: 99%

“…Results also show that in scale-free networks, a small proportion of consumers (hubs, or early adopters) may have an exceptional influence on the consumptive behavior of others. Rahmandad and Sterman (2008), while primarily concerned with comparing stochastic agent-based and deterministic differential equation models, also study the impact of different network structures. In particular, they compare fully connected, random, small-world, scale-free and lattice networks.…”

Section: Small-world Versus Regular Versus Random Network a Number Omentioning

confidence: 99%

Agent-based simulation of innovation diffusion: a review

Kiesling

Günther

Stummer

et al. 2011

Cent Eur J Oper Res

430

243

View full text Add to dashboard Cite

Mathematical modeling of innovation diffusion has attracted strong academic interest since the early 1960s. Traditional diffusion models have aimed at empirical generalizations and hence describe the spread of new products parsimoniously at the market level. More recently, agent-based modeling and simulation has increasingly been adopted since it operates on the individual level and, thus, can capture complex emergent phenomena highly relevant in diffusion research. Agent-based methods have been applied in this context both as intuition aids that facilitate theory-building and as tools to analyze real-world scenarios, support management decisions and obtain policy recommendations. This review addresses both streams of research. We critically examine the strengths and limitations of agent-based modeling in the context of innovation diffusion, discuss new insights agent-based models have provided, and outline promising opportunities for future research. The target audience of the paper includes both researchers in marketing interested in new findings from the agent-based modeling literature and researchers who intend to implement agent-based models for their own research endeavors. Accordingly, we also cover pivotal modeling aspects in 123 184 E. Kiesling et al. depth (concerning, e.g., consumer adoption behavior and social influence) and outline existing models in sufficient detail to provide a proper entry point for researchers new to the field.

show abstract

“… The cognitive limitations of humans such as the number of factors that can be considered in parallel, the amount of information that can be processed and the speed of processing (Bood & Postma, 1997;Oehmen et al, 2015).  The limited time to make a decision (Oehmen et al, 2015;Rahmandad & Sterman, 2008).…”

Section: Uncertaintymentioning

confidence: 99%