Cluster analyses of Lunar-Mars life support test project simulations

Ryan P. Goss, Haibei Jiang, Luis F Rodriguez

Research output: Contribution to conferencePaper

Abstract

BioSim, a life support system modeling tool has been utilized to analyze several scenarios considering a closed human-crop system. These scenarios are based on the Lunar-Mars Life Support Test Project (LMLSTP) conducted at NASA-Johnson Space Center. Initially, a single crew member was integrated within a vacuum pressure chamber which contained an 11.2 m 2 growth chamber. Recently, several heuristic and procedural tools were designed for optimization of scenarios based on the original NASA project. The simulation output was expected to be a scattered distribution but in actuality, several distinct tiers of output were observed. This tiering effect led to the hypothesis that certain crop subsystem configurations inherently perform very well, while others perform poorly. The objective of this work has been to identify the characteristics of these configurations and determine whether any correlation exists with the quality of the system. Four statistical clustering analysis procedures were selected and implemented in SAS (SAS Institute, Cary, NC) to identify the characteristics of the tiers found within the analysis output. The cluster analysis procedure has been tested on preliminary simulation results. Studying the analyses yielded clearly identifiable clusters for many of the different scenarios. However, in one of our more complex scenarios, the clustering procedure gave conflicting results among the different methods. We attribute these conflicting results to an empirically developed stochastic approach for determining system failure. Consequently, a deterministic approach to system failure has been implemented. Analysis of this approach is in progress.

Original languageEnglish (US)
StatePublished - Nov 7 2007
Event2007 ASABE Annual International Meeting, Technical Papers - Minneapolis, MN, United States
Duration: Jun 17 2007Jun 20 2007

Other

Other2007 ASABE Annual International Meeting, Technical Papers
CountryUnited States
CityMinneapolis, MN
Period6/17/076/20/07

Fingerprint

Mars
Crops
Cluster Analysis
NASA
United States National Aeronautics and Space Administration
support systems
Cluster analysis
crops
growth chambers
cluster analysis
Life Support Systems
testing
Vacuum
Pressure
Growth
methodology

Keywords

  • Advanced life support system
  • Cluster analysis
  • Genetic algorithm
  • Heuristics
  • Nonlinear regression
  • System optimization

ASJC Scopus subject areas

  • Agricultural and Biological Sciences(all)
  • Engineering(all)

Cite this

Goss, R. P., Jiang, H., & Rodriguez, L. F. (2007). Cluster analyses of Lunar-Mars life support test project simulations. Paper presented at 2007 ASABE Annual International Meeting, Technical Papers, Minneapolis, MN, United States.

Cluster analyses of Lunar-Mars life support test project simulations. / Goss, Ryan P.; Jiang, Haibei; Rodriguez, Luis F.

2007. Paper presented at 2007 ASABE Annual International Meeting, Technical Papers, Minneapolis, MN, United States.

Research output: Contribution to conferencePaper

Goss, RP, Jiang, H & Rodriguez, LF 2007, 'Cluster analyses of Lunar-Mars life support test project simulations' Paper presented at 2007 ASABE Annual International Meeting, Technical Papers, Minneapolis, MN, United States, 6/17/07 - 6/20/07, .
Goss RP, Jiang H, Rodriguez LF. Cluster analyses of Lunar-Mars life support test project simulations. 2007. Paper presented at 2007 ASABE Annual International Meeting, Technical Papers, Minneapolis, MN, United States.
Goss, Ryan P. ; Jiang, Haibei ; Rodriguez, Luis F. / Cluster analyses of Lunar-Mars life support test project simulations. Paper presented at 2007 ASABE Annual International Meeting, Technical Papers, Minneapolis, MN, United States.
@conference{8acf74a8c88e48c08fc2e0f0a27525e2,
title = "Cluster analyses of Lunar-Mars life support test project simulations",
abstract = "BioSim, a life support system modeling tool has been utilized to analyze several scenarios considering a closed human-crop system. These scenarios are based on the Lunar-Mars Life Support Test Project (LMLSTP) conducted at NASA-Johnson Space Center. Initially, a single crew member was integrated within a vacuum pressure chamber which contained an 11.2 m 2 growth chamber. Recently, several heuristic and procedural tools were designed for optimization of scenarios based on the original NASA project. The simulation output was expected to be a scattered distribution but in actuality, several distinct tiers of output were observed. This tiering effect led to the hypothesis that certain crop subsystem configurations inherently perform very well, while others perform poorly. The objective of this work has been to identify the characteristics of these configurations and determine whether any correlation exists with the quality of the system. Four statistical clustering analysis procedures were selected and implemented in SAS (SAS Institute, Cary, NC) to identify the characteristics of the tiers found within the analysis output. The cluster analysis procedure has been tested on preliminary simulation results. Studying the analyses yielded clearly identifiable clusters for many of the different scenarios. However, in one of our more complex scenarios, the clustering procedure gave conflicting results among the different methods. We attribute these conflicting results to an empirically developed stochastic approach for determining system failure. Consequently, a deterministic approach to system failure has been implemented. Analysis of this approach is in progress.",
keywords = "Advanced life support system, Cluster analysis, Genetic algorithm, Heuristics, Nonlinear regression, System optimization",
author = "Goss, {Ryan P.} and Haibei Jiang and Rodriguez, {Luis F}",
year = "2007",
month = "11",
day = "7",
language = "English (US)",
note = "2007 ASABE Annual International Meeting, Technical Papers ; Conference date: 17-06-2007 Through 20-06-2007",

}

TY - CONF

T1 - Cluster analyses of Lunar-Mars life support test project simulations

AU - Goss, Ryan P.

AU - Jiang, Haibei

AU - Rodriguez, Luis F

PY - 2007/11/7

Y1 - 2007/11/7

N2 - BioSim, a life support system modeling tool has been utilized to analyze several scenarios considering a closed human-crop system. These scenarios are based on the Lunar-Mars Life Support Test Project (LMLSTP) conducted at NASA-Johnson Space Center. Initially, a single crew member was integrated within a vacuum pressure chamber which contained an 11.2 m 2 growth chamber. Recently, several heuristic and procedural tools were designed for optimization of scenarios based on the original NASA project. The simulation output was expected to be a scattered distribution but in actuality, several distinct tiers of output were observed. This tiering effect led to the hypothesis that certain crop subsystem configurations inherently perform very well, while others perform poorly. The objective of this work has been to identify the characteristics of these configurations and determine whether any correlation exists with the quality of the system. Four statistical clustering analysis procedures were selected and implemented in SAS (SAS Institute, Cary, NC) to identify the characteristics of the tiers found within the analysis output. The cluster analysis procedure has been tested on preliminary simulation results. Studying the analyses yielded clearly identifiable clusters for many of the different scenarios. However, in one of our more complex scenarios, the clustering procedure gave conflicting results among the different methods. We attribute these conflicting results to an empirically developed stochastic approach for determining system failure. Consequently, a deterministic approach to system failure has been implemented. Analysis of this approach is in progress.

AB - BioSim, a life support system modeling tool has been utilized to analyze several scenarios considering a closed human-crop system. These scenarios are based on the Lunar-Mars Life Support Test Project (LMLSTP) conducted at NASA-Johnson Space Center. Initially, a single crew member was integrated within a vacuum pressure chamber which contained an 11.2 m 2 growth chamber. Recently, several heuristic and procedural tools were designed for optimization of scenarios based on the original NASA project. The simulation output was expected to be a scattered distribution but in actuality, several distinct tiers of output were observed. This tiering effect led to the hypothesis that certain crop subsystem configurations inherently perform very well, while others perform poorly. The objective of this work has been to identify the characteristics of these configurations and determine whether any correlation exists with the quality of the system. Four statistical clustering analysis procedures were selected and implemented in SAS (SAS Institute, Cary, NC) to identify the characteristics of the tiers found within the analysis output. The cluster analysis procedure has been tested on preliminary simulation results. Studying the analyses yielded clearly identifiable clusters for many of the different scenarios. However, in one of our more complex scenarios, the clustering procedure gave conflicting results among the different methods. We attribute these conflicting results to an empirically developed stochastic approach for determining system failure. Consequently, a deterministic approach to system failure has been implemented. Analysis of this approach is in progress.

KW - Advanced life support system

KW - Cluster analysis

KW - Genetic algorithm

KW - Heuristics

KW - Nonlinear regression

KW - System optimization

UR - http://www.scopus.com/inward/record.url?scp=35648941800&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=35648941800&partnerID=8YFLogxK

M3 - Paper

AN - SCOPUS:35648941800

ER -