Dynamic Multi-Stage Modeling for Commercial Aircraft Contracting Statement of Work: Air Mobility Command uses commercial aircraft contracts to supplement and augment military airlift capacity. When making decisions on contracting, numerous factors need to be considered including (1) capacity (2) contract execution and cancellation (3) unanticipated events. Thus, a dynamic decision model with high fidelity is of importance. Specifically, this model should: 1. provide better forecast on demand 2. provide insights on pricing mechanism for contract 3. assist reliable decision making on airlift contracting with anticipated demand changes 4. assist robust decision making with respect to unanticipated events 5. increase overall knowledge of the system and its level of performance 6. assess different recommendation scenarios for potentially increasing the cost-effectiveness of commercial aircraft acquisition To address the broad spectrum of the problem, a full range of technologies and methodologies must be synergistically employed. These include: Principle of stochastic optimization Use of constructive simulation Principles of systems engineering to include building early assessment influence Principles of reliability, maintainability and usability; and Integrated developmental and operational planning coupled with full requirements traceability In order to support the larger AFIT research effort, this research proposal intends to develop a generalized mathematical framework for resource/task allocations in a dynamic environment. The outcome will provide a scalable and timely algorithm to support current and future allocation decisions and one application is the commercial aircraft contracting. The project is being conducted in five general phases over three years: Phase I: Develop decentralized negotiation algorithm for dynamic pricing (January 2012-May 2012). Will be completed on time based on the current contract PR F4F5AT1165A001. Marketbased negotiation algorithm was developed (a paradigm from Distributed Artificial Intelligent and Multi-agent system) using linear-price and non-linear price for converged solutions. An accurate forecasting is needed for this phase. TASK: Phase II - Develop single-stage probability based optimization model including chance constraint, robust optimization for allocation decision (Aug 2012- May 2013). Other than traditional models from operational research, models from engineering design (e.g., reliability based design optimization, robust design optimization) will be studied. We will identify the appropriate models for the commercial aircraft contracting problem. After Phase II (Aug 2012 May 2013) appropriate models have been identified, research findings will be documented and conference papers will be made to be submitted to ISERC and/or INFORMS conference for peer review. In addition, a journal paper should be ready to be submitted by Jan 2013 to provide additional peer feedback and archive the research. Some targeted journals should include Computers and Operations Research and the European Journal of Operational Research. A review of Phase II research will be conducted by AFIT and the sponsor to determine if any future research is necessary. Future research would include the following phases. Phase III: Develop multi-stage decisions model (June 2013 Aug 2013). As the civilian commercial aircraft contracting decision not only has a yearly component but also a quarterly and monthly decision, it may be necessary to extend the models from Phase II and reformulate them into a multi-stage decision model. Phase IV: Develop simulation model as a testbed to validate Phase II and Phase III outcomes (Aug 2013 May 2014). Based on the historical data, development of a discrete event simulation model may be necessary to validate phase II and Phase III probabilistic models. The model is used to test the performance of the decision models from earlier phases. Phase IV: Integration (May 2014-May 2015). The final phase of this effort involves bringing all four phases of the research into an operable model that can react to the environment and transition to new optimal solutions smoothly without disrupting operations.
|Effective start/end date||8/1/12 → 5/31/13|
- DOD-USAF: Air Force Institute of Technology (AFIT): $45,000.00
Discrete event simulation
Multi agent systems