Building occupancy simulation and data assimilation using a graph-based agent-oriented model

Rai, Sanish; Hu, Xiaolin

doi:10.1016/j.physa.2018.02.051

Cited by 8 publications

(7 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As we have discussed above, extending Kalman filter based methods to ABMs requires a microstate synthesis algorithm adapted to the particular data structures present in the ABM. Generating such an algorithm has been straightforward in prior work ( 23, 24 ) because of the presence of discrete model compartments which house agents; these compartments allow for the macrostate to be composed of only agent counts per compartment. However, the presence of such compartments is not guaranteed in general, especially in ABMs which feature spatial attributes.…”

Section: Resultsmentioning

confidence: 99%

“…Validated computational models for these disease processes will often include complex, hybrid, agent-based, multi-scale, and stochastic dynamics. While there has extensive work on applying data assimilation (DA) techniques to these types of models (5)(6)(7), especially in the context of transportation systems (8)(9)(10)(11)(12)(13)(14)(15)(16)(17)(18) and social/epidemiological spread (19)(20)(21)(22)(23)(24), there is little work which has been done on biomedical models. Moreover, the work that has been done generally falls into two categories: (i) DA done on the entire description of an ABM, and (ii) DA done on surrogate models built from ABMs.…”

Section: Introductionmentioning

confidence: 99%

“…We intend to perform data assimilation using Kalman filter methods on macrostates, as implied in ( 23 ) and fully proposed in ( 24 ). However, there is an immediate problem: our models require the full microstate to run, and while the KF gives us a method of updating the distribution of model macrostates after an observation, it does not immediately lead to methods whereby we can update the corresponding distribution of microstates.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Ensemble Kalman filter methods for agent-based medical digital twins

Knapp,

Cruz,

Mehrad

et al. 2024

Preprint

View full text Add to dashboard Cite

A medical digital twin is a computational replica of some aspect of a patient’s biology relevant to patient health. It consists of a computational model that is calibrated to the patient and is dynamically updated using a stream of patient data. The underlying computational model is often, multi-state, stochastic, and combines different modeling platforms at different scales. Standard methods for data assimilation do not directly apply to such models. This is true in particular for ensemble Kalman filter (1, 2) methods, a common approach to such problems. This paper focuses on agent-based models (ABMs), a model type often used in biomedicine. The key challenge for any forecasting algorithm for this model type is to bridge the gap between (detailed) micro- and (summary) macrostates. This paper proposes a modified Kalman filter method to meet this challenge, providing a way to dynamically update the microstate of an ABM using patient measurements collected at the macro level.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Ensemble Kalman filter methods for agent-based medical digital twins

Knapp,

Cruz,

Mehrad

et al. 2024

Preprint

View full text Add to dashboard Cite

show abstract

“…Graph-based building simulation and data assimilation [32] 2018 Occupancy Agent-based Airport terminal…”

Section: School Buildingmentioning

confidence: 99%

Understanding Occupant Behaviors in Dynamic Environments using OBiDE framework

Arslan¹,

Cruz²,

Ginhac

2019

Building and Environment

View full text Add to dashboard Cite

Occupants' movements and presence are fundamental and the prerequisites for any type of occupant behaviors' understanding which tells whether a building location is occupied, the number of occupants or an occupant with a specific profile in a certain location. Numerous studies have been conducted over the past few decades to model occupant behaviors stochastically for an improved understanding of their activities for different facility management applications. Despite many research efforts to model dynamic behaviors of building occupants, the understanding of their behaviors by incorporating the contextual information linked to the evolution of the building environment is still not adequately explored. The contextual information linked to locations in dynamic environments changes often over time in terms of position, size, properties and relationships with the environment. This changing building environment affects the occupants' movements and presence inside the facility which ultimately degrades the process of inferring their accurate activities based on the location context. Henceforth, the evolving building information is required to be mapped with occupant movements for an improved understanding of their changing behaviors. To fill this research gap, a framework named 'Occupant Behaviors in Dynamic Environments' (OBiDE) is designed for providing a 'blueprint map' to integrate existing DNAS (Drivers, Needs, Systems, Actions) ontology (i.e. a scheme to model occupant behaviors) with our semantic trajectory enrichment model to better understand the occupant behaviors by tracking the dynamicity of building locations. The proposed framework extends the usability of DNAS by providing a centralized knowledge base that holds the movements of occupants with relevant historicized contextual information of the building environment to study occupant behaviors for different facility management applications.

show abstract

“…To address the limitations of the models and take advantage of the good features of both models, we developed a new model called graph-based agent-oriented model [3] that combines the better property of both models, i.e. scalability and diversity.…”

Section: Introductionmentioning

confidence: 99%

Parallelization of Sequential Monte Carlo Methods in Building Occupancy Simulation and Data Assimilation

Rai

2019 ASEE Annual Conference &Amp; Exposition Proceedings

Self Cite

View full text Add to dashboard Cite

Building occupancy models for large building and occupancy is a computationally costly simulation. Data can be collected about the building from various sensors installed in the building and can be used for realtime estimation modelling. The simulation model and real-time estimation algorithm both are complex models requiring resource to accommodate for increase in size of the building environment, its occupants and their activities. In our previous work, we presented a graph-based agent-oriented model to efficiently simulate large number of occupants and used Sequential Monte Carlo (SMC) methods to assimilate the occupancy data for real-time estimation. A main issue with SMC algorithm is its high computational cost, which increases proportionally with the simulation model size. To manage the need of resource when increasing particles, we implement parallelization to SMC and improve the runtime of overall building occupancy simulation and estimation.

show abstract

Building occupancy simulation and data assimilation using a graph-based agent-oriented model

Cited by 8 publications

References 10 publications

Ensemble Kalman filter methods for agent-based medical digital twins

Ensemble Kalman filter methods for agent-based medical digital twins

Understanding Occupant Behaviors in Dynamic Environments using OBiDE framework

Parallelization of Sequential Monte Carlo Methods in Building Occupancy Simulation and Data Assimilation

Contact Info

Product

Resources

About