Area of Interest: OSW Cost Assessment and Program Evaluation (CAPE) Modeling and Simulation

DEPT OF DEFENSE · WASHINGTON HEADQUARTERS SERVICES (WHS) · WASHINGTON HEADQUARTERS SERVICES is soliciting Area of Interest: OSW Cost Assessment and Program Evaluation (CAPE) Modeling and Simulation. Solicitation number HQ003427AOI5. Classified under NAICS 541715 — Research and Development in the Physical, Engineering, and Life Sciences (except Nanotechnology and Biotechnology). Product/service code AC32 — R&D- DEFENSE SYSTEM: SHIPS (APPLIED RESEARCH/EXPLORATORY DEVELOPMENT). Responses are due by 12 November 2026 21:00 UTC. Posted on SAM.gov, the US federal contract opportunities system; registration is required to bid.

The Department of War (DOW) Cost Assessment and Program Evaluation (CAPE) supports the Office of the Secretary of War (OSW). The Director of CAPE is the principal staff assistant to the Secretary of War for Cost Assessment & Program Evaluation. In fulfilling this role, the Director of CAPE provides independent, non-advocate, and data-driven analytical evaluations directly to the Secretary and Deputy Secretary of War regarding defense program investments, major force structure posture, and strategic resourcing trade-offs. The Director’s principal responsibilities include:

• Analyze and evaluate plans, programs, and budgets in relation to U.S. defense objectives, projected threats, allied contributions, estimated costs, and resource constraints.
• Review, analyze, and evaluate programs, including classified programs, for executing approved policies.
• Provide leadership in developing and promoting improved analytical tools and methods for analyzing national security planning and the allocation of resources.
• Ensure that the costs of DoW programs, including classified programs, are presented accurately and completely.
• Assess effects of DoW spending on the U.S. economy and evaluate alternative policies to ensure that DoW programs can be implemented efficiently.

Currently, CAPE’s campaign- and mission-level Modeling and Simulation (M&S) relies on heavy, time-intensive physics engines (e.g., AFSIM) that execute discrete, bespoke combat runs. While highly detailed, these simulation approaches are too slow, computationally expensive, and inflexible to support executive-level sensitivity sweeps across multi-dimensional threat environments and Department warfighting capabilities. CAPE’s current environment leverages HPC resources that provide 1.5 terabytes (TB) of usable memory and 2,000 Central Processing Unit (CPU) cores with no wall clock time restraints. On average, individual simulation execution via AFSIM requires 1 hour of real time to complete and 20 gigabytes (GB) of accessible RAM. This constraint, coupled with sharp disparities in compute architecture, hardware availability, and security accreditation across existing CAPE enclaves restrict analysts to limited, manual workflows while preventing the execution of high-throughput, compute-intensive campaign analytics. CAPE is seeking innovative solutions to establish a next-generation modeling architecture, replacing its legacy M&S infrastructure. To dramatically compress the cycle time between defense budget queries and statistically defensible warfighting insights, the target environment must integrate three interdependent functional layers: (1) A high-performance simulation execution layer to execute large numbers of model (e.g., AFSIM) runs to span the parameter space. Runs in this layer are computationally intensive and resource demanding but will be executed infrequently and systematically using advanced Design of Experiments techniques. (2) A machine-learned surrogate emulation layer to train ultra-fast emulator models from simulation results and detect out-of-sample queries. While the government plans to develop and calibrate these surrogate emulators and analytical models in-house, the commercial solution must provide the extensible data staging, pipeline integration, and execution framework to ingest batch simulation results and host these government-developed models. (3) An intuitive self-service quantitative analysis layer housing pre-built methods for sensitivity analysis, marginal gain estimates, etc., and tools for custom analysis (e.g., tradespace decisions between range, speed, and lethality of weapon systems) and visualization. The solution must support both current operational needs and have flexibility to accommodate future growth in data volume, model complexity, user demand, and computational intensity. This next generation environment will establish an open, scalable, and unified computational backbone designed to modernize Department decision support. The scope of this effort includes the following core technical and performance requirements.

• A high-performance compute environment sized to meet demanding modeling, simulation, and analytic workloads. Proposed solution architectures should span options to allow for parameter evaluation in the simulation layer on the order of 1,200 individual simulations (i.e., light case - isolated missions only) to be executed within 10 hours real-time, up to 150,000 individual simulations (i.e., heavy case – full joint campaign) to be executed within 2 weeks real-time. Of note, current CAPE architecture would require 1 year of simulation execution runtime for the heavy case.
• Secure, efficient, data and software movement both into and out of the environment, as well as the ability to interface with external networks to enable transfer of data in accordance with applicable security and access controls. The environment must support network throughput of 100 Mbps - 1 Gbps. The solution must also support connectivity over HTTPS to designated external OSW and mission-relevant data sources, including OSW CAPE-furnished databases and other authorized repositories (e.g., War Data Platform), as applicable.
• Ability to accommodate evolving development ecosystems and diverse sets of software including M&S tools and analytic technologies. This includes modern software (e.g., Python, R, Java, etc.), scalable data platforms, scalable parallelizable processes (e.g., containers), optimization tools (e.g., GAMS), artificial intelligence capabilities (e.g., large language models), and neural network development (e.g., PyTorch), provided they can be incorporated in a secure and maintainable manner. The solution must enable efficient onboarding, integration, and removal of models and software as mission needs evolve.
• The Government will retain full operational tenant administration, maintaining exclusive control over user access provisioning, role-based security permissions, data ingestion, proprietary model uploads, container orchestration, and analytical query execution. The Commercial provider will be responsible for turnkey, managed platform and infrastructure IT operations, including hardware availability, operating system patching, network performance, automated system backups, and baseline cybersecurity hardening, ensuring government analysts and technical leads are fully insulated from day-to-day hardware maintenance and IT administration.