No description available.

No description available.

No description available.

The application of a mineral admixture, or a corrosion inhibitor, or a combination of both are methods used for the corrosion protection for reinforced concrete bridges. The results of a study on evaluation of corrosion inhibitors from three different manufacturers and of concretes with fly ash, slag cement, and silica fume and a concrete with silica fume and a corrosion inhibitor are presented. The specimens were built to simulate four exposure conditions typical for concrete bridges located in the coastal region or inland where deicing salts are used. The exposure conditions were horizontal, vertical, tidal, and immersed zones. The specimens were kept inside the laboratory and were exposed to weekly ponding cycles of 6% (w/w) sodium chloride solution. In addition, cover depth measurements from 21 bridge decks and chloride data from 3 bridge decks were used, together with laboratory data, in modeling the service lives of the investigated corrosion protection methods. The methods used to assess the condition of the specimens included chloride concentration measurements, corrosion potentials, and corrosion rates (3LP). Additionally, visual observations were performed for identification of rust stains and cracking on concrete surfaces. Modeling the time as a function of probability of the end of functional service life (EFSL) is presented. It has been shown that the distributions of surface chloride concentration, C0, and diffusion coefficient, Dc, are key elements in the model. Model predictions show that the concretes with mineral admixtures provide much better level of protection against moisture and chlorides than the ordinary portland cement concrete alone. Application of a corrosion inhibitor causes an elevation of the chloride threshold resulting in an additional increase in time to EFSL. More field studies are needed to better estimate distributions of surface chloride concentration and diffusion coefficient of Virginia bridge decks, and to confirm predicted times to EFSL for low permeable (LP) concretes. / Ph. D.

A functional analysis, conducted to assess the variables maintaining the chronic eye poking of a female diagnosed with profound mental retardation, indicated that the behavior persisted in the absence of social contingencies. A procedure was initiated in a training environment in which a punisher (mild reprimand) was delivered contingent on eye poking in the presence, but not in the absence, of a neutral stimulus (wristbands). Using a combination of multiple baseline and multielement experimental designs, it was determined that that eye poking was suppressed in the presence of the previously neutral stimulus, even in environments in which the reprimand contingency was inoperative.

Research shows that human operant behavior typically differs from non-human operant behavior on schedules of reinforcement. These differences in performance may be related to differences between the experimental preparations used to study human and non-human operant behavior. One such difference is the type of reinforcer used. This experiment analyzed the differential effects of points alone, points backed up by money, and sounds on schedule performance of human subjects. Results show that sounds generated moderate rates of responding, capable of change in either direction. When points backed up with money were the reinforcers, however, high rates of behavior were generated, disrupting the previously established baseline performance. This suggests that while points may be effective in generating high rates of behavior, they may be ineffective in producing sensitive baselines needed to study human operant behavior on schedules of reinforcement.

Historically, horse trainers have relied primarily upon repetition, negative reinforcement, and punishment to teach new behaviors. Positive reinforcement has been eschewed, largely on the basis of the wides read belief that positive reinforcement is not effective with horses. Additional difficulties in the timely application of such reinforcement have further inhibited its use. After repeated pairing of an auditory stimulus with an established primary reinforcer, the auditory stimulus was predicted to be a reinforcer. An equine subject was then successfully trained to perform five different, novel tasks using only the auditory stimulus. Subsequently, extinction of behavior was noted in the absence of the conditioned reinforcer. Implications for many phases of horse training were discussed. Some weaknesses of the present study were noted along with suggested issues for future investigations.

The purpose of this study was to develop an effective method for trailer loading horses based on the principles of positive reinforcement. Target training and shaping were used to teach trailer loading behavior in a natural setting. Five AQHA mares were selected for this program. All five had been loaded before through the use of punishment. A two-horse trailer was used. Approximations to loading and inappropriate behaviors were the dependent variables. When intervention started the target was moved to various locations inside the trailer. Subjects started training on the left side of the trailer. After a subject was loading in the left side they were moved to the right side, then to loading half on the right and half on the left, then they were loaded by a different trainer, and into a different trailer. For one subject a limited hold was utilized, as well as a companion horse.

The study examined the effects of positive verbal reinforcement on intrinsic motivation by determining differential effects over four multidimensions of Ryan's Intrinsic Motivation Inventory (IMI). Subjects (N=60) were 30 male and 30 female college students. The subjects were blocked by gender and randomly assigned to a positive verbal reinforcement group or a control group. The subjects received 10 trials on the stabilometer. The results of the study indicated that there were significant group differences for composite intrinsic motivation and for perceived competence; however, there were no significant gender differences found. Furthermore, no group differences were reported for the underlying factors of interest/enjoyment, effort, or pressure/tension.

"This thesis focuses on two key problems in reinforcement learning: How to design reward functions to obtain intended behaviors in autonomous systems using the learning-based control? Given complex mission specification, how to shape the reward function to achieve fast convergence and reduce sample complexity while learning the optimal policy? To answer these questions, the first part of this thesis investigates inverse reinforcement learning (IRL) method with a purpose of learning a reward function from expert demonstrations. However, existing algorithms often assume that the expert demonstrations are generated by the same reward function. Such an assumption may be invalid as one may need to aggregate data from multiple experts to obtain a sufficient set of demonstrations. In the first and the major part of the thesis, we develop a novel method, called Non-parametric Behavior Clustering IRL. This algorithm allows one to simultaneously cluster behaviors while learning their reward functions from demonstrations that are generated from more than one expert/behavior. Our approach is built upon the expectation-maximization formulation and non-parametric clustering in the IRL setting. We apply the algorithm to learn, from driving demonstrations, multiple driver behaviors (e.g., aggressive vs. evasive driving behaviors). In the second task, we study whether reinforcement learning can be used to generate complex behaviors specified in formal logic — Linear Temporal Logic (LTL). Such LTL tasks may specify temporally extended goals, safety, surveillance, and reactive behaviors in a dynamic environment. We introduce reward shaping under LTL constraints to improve the rate of convergence in learning the optimal and probably correct policies. Our approach exploits the relation between reward shaping and actor-critic methods for speeding up the convergence and, as a consequence, reducing training samples. We integrate compositional reasoning in formal methods with actor-critic reinforcement learning algorithms to initialize a heuristic value function for reward shaping. This initialization can direct the agent towards efficient planning subject to more complex behavior specifications in LTL. The investigation takes the initial step to integrate machine learning with formal methods and contributes to building highly autonomous and self-adaptive robots under complex missions."