FRANK W. MOORE

1791 Shadylane Dr., Beavercreek, OH 45432-2064

(937) 427-1534

fmoore@valhalla.cs.wright.edu

Citizenship: USA

 

EDUCATION

 

BSCE, Computer Engineering (3/86)

MSCE, Computer Engineering (3/88)

PhD, Computer Science & Engineering (8/97)

Wright State University (WSU), Dayton, OH

 

DISSERTATION: "A Methodology for Strategy Optimization Under Uncertainty". Defined a methodology for using genetic programming to solve complex strategy optimization problems (including Missile Countermeasures Optimization) under conditions of uncertainty. Results presented at GP-97, NAECON-97, and MAICS-97. See attached poster paper for description.

 

MASTERS THESIS: "A Dual-Processor Shared Memory System". Designed and implemented dual-port random-access memory boards and a dual-processor shared memory operating system kernel. Demonstrated the advantages of this system for solving such classic multi-tasking problems as bi-directional heuristic search.

 

Awarded Dayton-Area Graduate Studies Institute (DAGSI) competitive scholarships in 1995-96 and 1996-97. Also received an Ohio Board of Regents Scholarship (1980-84), and various research and teaching assistantships at WSU, as described below.

 

EXPERIENCE

 

RESEARCH GRANT, WSU (9/95-3/96): As knowledge engineer, worked closely with an expert pilot and human-factors analysts to successfully implement the Approach Procedures Expert System (APES), an embedded knowledge-based decision-aiding expert system intended to help pilots fly approaches. APES software used voice and textual messages to guide the pilot through the proper sequence of tasks necessary to fly the current approach, while continuously monitoring aircraft performance and notifying the pilot of proper corrective actions. The APES received high marks from all 15 test subjects, and subjective workload analysis demonstrated significant improvement in pilot performance, particularly under high-workload conditions without prior review of the approach plate. Research supported by the Flight Dynamics Directorate, Wright-Patterson Air Force Base (W-PAFB), OH. Results presented at NAECON-96 and HICS-96.

 

SENIOR SYSTEMS ENGINEER, Sumaria Systems, Inc. (7/93-8/95). Determined a methodology for weapons systems requirements specification, traceability, and analysis in support of the RAPID-WS program at the Armstrong Laboratory Human Systems Division, W-PAFB, OH. Also provided software support for the Electronic Commerce Resource Center (ECRC) in Dayton, OH. Received the first Dayton ECRC Employee of the Month award, as well as a letter of commendation from the Defense Electronics Supply Center (DESC) commander for EC/EDI training provided to DESC employees.

 

MTS II, Veda Inc. (1/92-3/93). Developed, modified, and maintained numerous simulation and human factors assessment software models in support of the Crew-Centered Cockpit Design (CCCD) program at the Armstrong Laboratory Human Engineering Division, W-PAFB, OH. The CCCD program provided a test bench used to identify improvements in aircraft cockpit design that reduce overall pilot workload. Designed and implemented an improved Timeline Management Tool, allowing human factors experts to perform task decomposition and subsequent workload analysis.

 

MTS I, TRW Inc. (9/89-1/92). Developed, modified, and maintained numerous avionics software models in support of the Avionics Integration Support Technology (AIST) and Integrated Test Bed (ITB) programs at the Avionics Laboratory, W-PAFB, OH. These programs provided distributed real-time simulated test environments supporting the integration of various avionics models into high-performance aircraft. Received a letter of commendation from the ITB principal investigator in recognition of successful implementation of real-time data bus monitoring, simulated system mass memory, and mass memory image generation software. As part of a four-person team, converted 48 separate avionics models to Ada, isolated machine dependencies to a set of data and file type definition packages, and successfully rehosted the entire AIST model set in less than two days.

 

FACULTY MEMBER, WSU Department of Computer Science & Engineering: Instructor (9/87-8/89), Adjunct Instructor (9/89-6/95). Taught graduate and undergraduate courses in Computer Science (CS), Computer Engineering (CEG), and Electrical Engineering (EE). Received consistently high student evaluations of instructor performance. Also provided academic advising for CS/CEG majors.

 

CEG 260, EE 351/551: Digital Computer Hardware

CEG 320/520: Computer Organization & Assembly Language Programming

CEG 360/560, EE 451/651: Digital Systems Design

CS 142, 145, 240, 241, 242: Introductory programming sequence

 

GRADUATE RESEARCHER, Graduate Student Summer Support Program (GSSSP), W-PAFB (Summer 1987). Investigated the potential impact of new photonic materials on optical processing technologies. Showed how these materials and various types of lasers could be used to construct ultra-high-speed digital optical logic gates (including inversion), AD/DA converters, and massively-parallel DMAC multipliers. These efforts established subsequent US Air Force support for several optical computing research projects at the WSU Research Center.

 

TEACHING ASSISTANT, WSU Department of Computer Science & Engineering (9/84-6/87, 3/93-8/93, 6/96-8/96). Provided graduate and undergraduate laboratory instruction in Computer Science & Engineering.

 

CS 141/142: Introductory programming sequence

CEG 392/592: Microprocessor-Based Systems Design

CEG 453/653: Design of Computing Systems

CS 710: Artificial Intelligence

 

PROFESSIONAL AFFILIATIONS

 

Association of Computing Machinery, 3/86-present. (WSU Student Chapter chairman, 3/86-3/87).

 

IEEE & IEEE Computer Society, 3/87-present.

 

PUBLICATIONS

 

See attached list.

 

REFERENCES

 

See attached list.

 


A Genetic Programming Methodology for Strategy Optimization

Under Uncertainty

 

Frank W. Moore and Dr. Oscar N. Garcia

Department of Computer Science and Engineering

Wright State University

Dayton, OH 45435

fmoore@valhalla.cs.wright.edu, ogarcia@valhalla.cs.wright.edu

 

ABSTRACT

This poster paper summarizes ongoing dissertation research defining a genetic programming (GP) methodology for solving strategy optimization problems for intelligent, autonomous agents operating under conditions of uncertainty. This dissertation provides an overview of current state-of-the-art approaches to strategy optimization problems; defines the Missile Countermeasures Optimization (MCO) problem as an instance of a strategy optimization problem; describes various types and degrees of uncertainty that may be introduced into the MCO problem; and develops a new methodology for solving the MCO problem under conditions of uncertainty. The new methodology promises significant improvements over current control-theoretic and analytical approaches.

 

Phase I of this research establishes a GP methodology for solving the MCO problem. The Extended Two-Dimensional Pursuer/Evader (E2DPE) problem is an abstracted, simplified MCO problem that models an evading aircraft and a pursuing surface-launched anti-aircraft missile (SAM) as point masses that use thrusting and turning forces to control their trajectory across a plane. The pursuer uses proportional navigation to pursue and destroy the evader. GP is used to evolve control programs that optimize evader survivability.

Best-of-run programs evolved by GP systems frequently prove to be brittle when subsequently used in situations not explicitly

anticipated during program evolution. Phase II investigates the use of randomly-generated fitness cases in the E2DPE problem to help evolve best-of-run programs that are less brittle than programs evolved using fixed training populations.

Uncertainty introduces a degree of complexity that is difficult to model using traditional analytical methods. Best-of-run programs evolved for one type of pursuer generally exhibit sub-optimal performance when subsequently tested against other pursuer types. Phase III introduces uncertainty about the type of pursuer, and evolves programs that solve the E2DPE problem for pursuer populations described by a probability distribution over possible types. During many encounters, the state of the SAM (its relative displacement, velocity, and acceleration) may also be unknown or uncertain. Phase IV investigates the impact of uncertainty about the pursuer’s current state, evolves programs under various conditions of uncertainty, and establishes upper and lower bounds on the probable fitness of the resulting best-of-run programs. For many encounters, maneuvers alone are inadequate to successfully evade an incoming SAM. For this reason, modern aircraft combine chaff, flares, jamming, and other countermeasures with maneuvers to increase survivability. In general, the effectiveness of these countermeasures depends upon the state of the aircraft and SAM at the time the countermeasure is deployed. Phase V introduces the use of additional countermeasures into the E2DPE problem, and evolves programs that combine these countermeasures with maneuvers to optimize evader survivability.

This dissertation concludes during Phase VI by identifying ways of extending the GP methodology of Phases I-V to solve the three-dimensional MCO problem. The authors gratefully acknowledge the support of the Dayton Area Graduate Studies Institute.