SOLICITATION NOTICE
A -- ME Receptor Model Development
- Notice Date
- 2/2/2004
- Notice Type
- Solicitation Notice
- Contracting Office
- Environmental Protection Agency, Ord Service Center/Nheerl, Rtp Procurement Operations Division, Research Triangle Park, NC 27711
- ZIP Code
- 27711
- Solicitation Number
- RFQ-RT-04-00125
- Response Due
- 2/17/2004
- Archive Date
- 3/17/2004
- Point of Contact
- Point of Contact, Ardra Morgan-Kelly, Purchasing Agent, Phone (919) 541-3670
- E-Mail Address
-
Email your questions to U.S. Environmental Protection Agency
(morgan.ardra@epa.gov)
- Description
- NAICS Code: 541511 NAICS CODE: 541511 THIS IS A COMBINED SYNOPSIS/SOLICITATION FOR COMMERCIAL SERVICES PREPARED IN ACCORDANCE WITH THE FORMAT IN FAR SUBPART 12.6 AS SUPPLEMENTED WITH ADDITIONAL INFORMATION INCLUDED IN THIS NOTICE. THIS ANNOUNCEMENT CONSTITUTES THE ONLY SOLICITATION. QUOTES ARE BEING REQUESTED AND A WRITTEN SOLICITATION WILL NOT BE ISSUED. The associated North Industry Classification System (NAICS) is 541511 and the business size standard is $21.0 million. Full and open competitive procedures will be unutilized. A firm, fixed price purchase order using Simplified Acquisition Procedures is anticipated to result from the award of this solicitation. This procurement is for ME Receptor Model Development. STATEMENT OF WORK: BACKGROUND The Positive Matrix Factorization (PMF) receptor model is a multivariate receptor model that uses both measured concentrations and uncertainty estimates to generate both particulate matter (PM) source profiles and source contribution estimates. The PMF receptor model uses a least squares approach to perform the factor analysis and incorporates non-negativity constraints (Paatero, 1997). The PMF receptor model is implemented as either PMF2 or PMF3 based on the structure of the data with PMF2 used for a single sampling site, and PMF3 used for multiple sampling sites. The Multilinear Engine (ME) receptor model was also developed by Paatero (1999), which uses a more flexible algorithm and a script language similar to Fortran or C. PMF can be implemented through ME, and ME allows for the incorporation of source profile constraints and the use of supplementary variables such as weekday/weekend and wind direction (Paatero, 2002). The ability to incorporate additional supplementary information and the flexibility of the ME program has allowed it to be used for the evaluation of both ambient and personal exposure data (Hopke et al., 2003). Standardized pre- and post- processing procedures do not exist for the ME or PMF receptor models. In addition, the ME model does not calculate uncertainties. Lastly, there are few measures of goodness of fit of the model to the data. The lack of standardized pre- and post- processing diagnostics, calculated uncertainty estimates, and indicators of how well the model fits the data limits the use of this powerful model for receptor modeling by State and Local Air Pollution Agencies to support State Implementation Plans for Particulate Matter. SCOPE The contractor shall develop a Quality Assurance Project Plan (QAPP) for the ME model development. The contractor shall develop a document describing recommended data pre-processing and diagnostics for the ME receptor model. The contractor shall provide a stand-alone executable of PMF2 implemented through the ME receptor modeling program described in Paatero et al. 1999, which can be distributed by EPA at no charge to the user. This version of the program will be registered to the U.S. Environmental Protection Agency. The executable shall allow transfer of sampling concentration, sample uncertainty, and supplementary data (factors such as weekend/weekday) to the program and will output model diagnostics, source profiles, sample source contribution estimates, and factor coefficients. In addition, the contractor shall develop algorithms to estimate the uncertainty of the source profiles, sample source contribution estimates, and factor coefficients. The contractor shall prepare a document that describes the method used to calculate the uncertainties. The contractor shall add the uncertainty algorithm to the executable. As an optional task, the contractor shall develop advanced model diagnostics that indicate how well the data meets the assumptions underlying the ME receptor model and develop summary information that indicates how well the specified ME model fits the data. The intentions of such diagnostics and summary statistics is to ease the comparison of one ME solution to another, to aide the user in identifying poor solutions, and to aide the user in improving the model and hence the solution. Possible diagnostics include correlation coefficients between observed and predicted data, patterns in the correlations such as poor correlation during certain time periods, cross-validation errors, influence of non-representative observations, indicators of variables or time steps for which there is so little signal in the data that the variable or observation should be removed from the modeling, indicators of amount of rotational ambiguity in a solution, and information about how different the source profiles can be due to the rotational ambiguity. TASKS TO BE COMPLETED Task 1. Provide a Quality Assurance Project Plan (QAPP) for the ME model development. The contractor shall provide to the EPA a report describing the ME model development quality assurance procedures (see Guidance for Quality Assurance Project Plans for Modeling EPA QA/G-5M, EPA/240/R-02/007, Dec 2002; http://www.epa.gov/quality/qs-docs/g5m-final.pdf): a) describe and provide references for theories/techniques/methodologies that will be used. b) assess whether available algorithms, and other information to be used in Tasks 2 - 4 are adequate and sufficiently complete to develop the pre- and post processing diagnostics and algorithms, ME executable, and ME uncertainty estimates. c) define the criteria for assessing the quality and reliability of the algorithms and of the model that is being developed and applied (availability and results of quality assurance/quality control (QA/QC) data, etc.); d) report the results from the information quality and reliability assessments, and; e) assess and ensure that the model, and model results are accurate and complete, including a QA evaluation of the process and products. Task 2: ME Pre- and Post-Processing The contractor shall create a document consisting of recommendations for ME receptor model pre- and post- processing requirements and diagnostics. The contractor shall create an algorithm using MATLAB to pre-process data for the ME executable. Similarly, the contractor shall develop algorithms using MATLAB for post-processing the output data from the ME executable. The contractor shall provide the algorithms to EPA. The contractor shall provide an example particulate matter data set, pre-processing results, ME output (source profiles, source contributions, and diagnostics), and post-processing diagnostics/figures. Task 3: ME Executable The contractor shall develop a stand-alone version of PMF2 implemented through the ME model described by Paatero (1999). The contractor shall provide an executable that can input and output selected variables. The input variables shall at least include total mass (i.e. PM2.5), species concentrations, species uncertainties, three supplementary variables or factors (weekday/weekend, season, wind direction), and other required model parameters such as the matrix dimensions, number of factors, robust/non-robust, stabilizer, positive/negative outlier distance, and missing data limit. The option for use of 0,1, 2, or 3 supplementary variables shall be allowed by the executable. Output variables shall at least include source profile, sample source contributions, factor coefficient estimates, scaled residuals, and model diagnostics (number of iterations, Q, and variables required for post processing (Task 2). Task 4: ME Uncertainty Estimates The contractor shall develop algorithms to estimate the uncertainty in the source profiles, individual sample source contribution estimates, factor coefficients, and average source contribution estimates. The contractor shall document the algorithms in a format suitable for submission to Chemometrics and Intelligent Laboratory Systems (15 pages). The contractor shall add the uncertainty algorithms to the Task 3 executable and provide a revised ME executable. Output variables in the revised ME executable shall at least include source profile and uncertainties, sample source contribution and uncertainties, factor coefficient estimates and uncertainties, scaled residuals, and model diagnostics (number of iterations, Q, and variables required for post processing (Task 2)). Optional Task 5: Advanced Model Diagnostics The contractor shall create a document consisting of recommendations for advanced model diagnostics for the ME receptor model. The contractor shall develop algorithms using MATLAB for performing these diagnostics and shall provide these algorithms to EPA. The contractor shall demonstrate these diagnostics on the example particulate matter data set used in Task 2. The diagnostics for this task should build on those developed in Task 2. Additionally, the diagnostics should tie directly into the uncertainty estimates developed in Task 4 in that models with large uncertainty estimates likely should have model diagnostics indicative of poor model performance whereas models with small uncertainty estimates likely should have diagnostics indicative of a model that is appropriate and fits the data well. DELIVERABLES Task 1. Provide a Quality Assurance Project Plan (QAPP) for the ME model development Provide a QAPP (Microsoft Word format) describing the ME model development quality assurance procedures. Deliverable date: 45 days after the start of the performance period. Task 2: ME Pre- and Post-Processing Provide document (Microsoft Word format) describing recommendations for pre- and post-processing requirements for the ME receptor model, and provide the pre-processing and post-processing algorithms (MATLAB m-files). Deliverable date: 3 months after the start of the performance period. Task 3: ME Executable Provide an ME executable and report (Microsoft Word format) describing the ME executable (inputs, outputs, and diagnostics) based on an example data set. Deliverable date: 6 months after the start of the performance period. Task 4: ME Uncertainty Estimates Provide document (Microsoft Word format) describing uncertainty algorithms. Add the uncertainty algorithms to the Task 3 executable and provide the executable. Deliverable date: 12 months after the start of the performance period. Optional Task 5: Advanced Model Diagnostics Provide document (Microsoft Word format) describing recommendations for advanced model diagnostics for the ME receptor model and provide MATLAB algorithms as m-files that implement the diagnostics. Deliverable date: 4 months after decision made to pursue optional task.. SELECTION CRITERIA: The EPA will select a vendor based upon both cost and technical factors so that the Agency receives the best value for this purchase where Technical and Past Performance are of greater importance than Price. Selection will be based on the following criteria: Knowledge of the algorithms employed in the software tools Positive Matrix Factorization 2 (PMF2) and the Multi-Linear Engine 2 (ME2) and experience in programming with the ME script language. Demonstrated development and application of the ME receptor model described in Paatero (1999) to determine and quantify the sources of PM using innovative approaches (i.e. supplementary variables such as weekend/weekday). Demonstrated development and application of algorithms for quantifying the uncertainty in PMF2 and ME2 solutions. Past performance contact information regarding 3 efforts (contracts, task orders, purchase order etc.) of similar complexity and size performed within the last 3 years. The following FAR provisions shall apply to the solicitation: 52.203-6 Restriction on Subcontractor Sales to the Government, 52.219-8 Utilization of Small Business Concerns , 52.204-6 Data Universal Numbering System, 52.204-7 Central Contractor Registration, 52.212-1, Instructions to Offerors -- Commercial Items. All offerors are to include with their offer a completed copy of provision 52.212-3, Offeror Representations and Certifications--Commercial Items. The following FAR clauses apply to the acquisition: 52.212-4, Contract Terms and Conditions--Commercial Items. The following additional FAR clauses which are cited in Clause 52.212-5 are applicable to this acquisition: 52.222-21, Prohibition of Segregated Facilities ; 52.222-26, Equal Opportunity, 52.222-35, Affirmative Action for Special Disabled and Vietnam Era Veterans; 52.222-36 Disabled Veterans and Veterans of the Vietnam Era; 52.222-19, Child Labor-Cooperation with Authorities and Remedies; 52.222-51, Buy American Act--Balance of Payments Program; 52.232-34, Payment of Electronic Funds Transfer. COMMERCIAL BUY CLAUSES AND FORMS are provided for your convenience at EPA's website: http://www.epa.gov/oam/rtp-cmd. Please submit to Ardra Morgan-Kelly, @ (919) 541-4273(fax) or email: morgan-kelly.ardra@epa.gov phone:(919) 541-3670. All offers are due by Feb12, 2004 1:00pm, EST. Vendor must include a technical proposal and cost proposal. The cost proposal shall include a price per task.
- Record
- SN00514392-W 20040204/040202212346 (fbodaily.com)
- Source
-
FedBizOpps.gov Link to This Notice
(may not be valid after Archive Date)
| FSG Index | This Issue's Index | Today's FBO Daily Index Page |