Connie Loggia Ramsey

Learning Sequential Decision Rules Using Simulation Models and Competition

The problem of learning decision rules for sequential tasks is addressed, focusing on the problem of learning tactical decision rules from a simple flight simulator. The learning method relies on the notion of competition and employs genetic algorithms to search the space of decision policies. Several experiments are presented that address issues arising from differences between the simulation model on which learning occurs and the target environment on which the decision rules are ultimately tested.

Putting more genetics into genetic algorithms

The majority of current genetic algorithms (GAs), while inspired by natural evolutionary systems, are seldom viewed as biologically plausible models. This is not a criticism of GAs, but rather a reflection of choices made regarding the level of abstraction at which biological mechanisms are modeled, and a reflection of the more engineering-oriented goals of the evolutionary computation community. Understanding better and reducing this gap between GAs and genetics has been a central issue in an interdisciplinary project whose goal is to build GA-based computational models of viral evolution. The result is a system called Virtual Virus (VIV). VIV incorporates a number of more biologically plausible mechanisms, including a more flexible genotype-to-phenotype mapping. In VIV the genes are independent of position, and genomes can vary in length and may contain noncoding regions, as well as duplicative or competing genes. Initial computational studies with VIV have already revealed several emergent phenomena of both biological and computational interest. In the absence of any penalty based on genome length, VIV develops individuals with long genomes and also performs more poorly (from a problem-solving viewpoint) than when a length penalty is used. With a fixed linear length penalty, genome length tends to increase dramatically in the early phases of evolution and then decrease to a level based on the mutation rate. The plateau genome length (i.e., the average length of individuals in the final population) generally increases in response to an increase in the base mutation rate. When VIV converges, there tend to be many copies of good alternative genes within the individuals. We observed many instances of switching between active and inactive genes during the entire evolutionary process. These observations support the conclusion that noncoding regions serve as scratch space in which VIV can explore alternative gene values. These results represent a positive step in understanding how GAs might exploit more of the power and flexibility of biological evolution while simultaneously providing better tools for understanding evolving biological systems.

A comparative analysis of methods for expert systems

Abstract Given the current widespread interest in expert systems, it is important to examine the relative advantages and disadvantages of the various methods used to build them. In this paper we compare three important approaches to building decision aids implemented as expert systems: Bayesian classification, rule-based deduction, and frame-based abduction. Our critical analysis is based on a survey of previous studies comparing different methods used to build expert systems as well as our own collective experience over the last five years. The relative strengths and weaknesses of the different approaches are analysed, and situations in which each method is easy or difficult to use are identified.

An evaluation of expert systems for software engineering management

The development of four separate, prototype expert systems to aid in software engineering management is described. Given the values for certain metrics, these systems provide interpretations which explain any abnormal patterns of these values during the development of a software project. The four expert systems which solve the same problem, were built using two different approaches to knowledge acquisition, a bottom-up approach and a top-down approach and two different expert system methods, rule-based deduction and frame-based abduction. In a comparison to see which methods might better suit the needs of this field, it was found that the bottom-up approach led to better results that did the top-down approach, and the rule-based deduction systems using simple rules provided more complete and correct solutions than did the frame-based abduction systems. >

Genome Length as an Evolutionary Self-adaptation

There is increasing interest in evolutionary algorithms that have variable-length genomes and/or location independent genes. However, our understanding of such algorithms both theoretically and empirically is much less well developed than the more traditional fixed-length, fixed-location ones. Recent studies with VIV (Virtual Virus), a variable length, GA-based computational model of viral evolution, have revealed several emergent phenomena of both biological and computational interest. One interesting and somewhat surprising result is that the length of individuals in the population self-adapts in direct response to the mutation rate applied, so the GA adaptivelv strikes the balance it needs to successfully solve the problem. Over a broad range of mutation rates, genome length tends to increase dramatically in the early phases of evolution, and then decrease to a level based on the mutation rate. The plateau genome length (i.e., the average length of individuals in the final population) generally increases in response to an increase in the base mutation rate. Furthermore, the mutation operator rate and adapted length resulting in the best problem solving performance is about one mutation per individual. This is also the rate at which mutation generally occurs in biological systems, suggesting an optimal, or at least biologically plausible, balance of these operator rates. These results suggest that an important property of these algorithms is a considerable degree of self-adaptation.

BaRT: A Bayesian Reasoning Tool for Knowledge Based Systems

Abstract As the technology for building knowledge based systems has matured, important lessons have been learned about the relationship between the architecture of a system and the nature of the problems it is intended to solve. We are implementing a knowledge engineering tool called BART that is designed with these lessons in mind. BART is a Bayesian reasoning tool that makes belief networks and other probabilistic techniques available to knowledge engineers building classificatory problem solvers. BART has already been used to develop a decision aid for classifying ship images, and it is currently being used to manage uncertainty in systems concerned with analyzing intelligence reports. This paper discusses how state-of-the-art probabilistic methods fit naturally into a knowledge based approach to classificatory problem solving, and describes the current capabilities of BART.

Monitoring an Ada software development

Abstract : Ada evolved from a desire within the Department of Defense to have a standard language for the development of real-time and large scale systems. In addition to providing features needed by those types of systems, Ada supports structured programming, data abstraction, modularity, and information hiding. Research with these techniques indicates that their use should improve the quality of the software development process and its product. While, programmers who are most familiar with various assembly languages and FORTRAN may use structured programming, generally they are not familiar with the other concepts. The problems with training programmers in Ada and its associated design and programming methods and then redeveloping current systems in Ada is unknown. In order to understand the effect of using Ada, the University of Maryland and the General Electric Company began a joint project. The purpose of the project is to monitor the use of Ada in an industrial software development project. In particular, we identify areas of success and difficulty in learning and using Ada as both a design and coding language. Our results indicate where emphasis should be placed in Ada training and in the development of tools and techniques for use with Ada. We also identify metrics used to evaluate and predict the cost, quality, and maintainability of Ada programs. (Author)