A singular values approach in helicopter gas turbine engines flight testing analysis

The process of empirical models evaluation is at the core business of experimental flight-testing data analysis. Accurate and convenient flight-testing of helicopter engine(s) available power is crucial for predicting the total helicopter performance. Common practice in estimation of in-flight helicopter gas turbine engine power consists of a reduction of flight-test data into simplistic single-variable analysis approach. While such an approach is convenient for practical use, it often results in unrealistic predictions of the available engine(s) power. A novel approach for the helicopter available power problem is the so-called Multivariable Polynomial Optimization under Constraints method. In this method, 18 regressors, constructed from the engine non-dimensional parameters, are used to define empirical polynomial models. This paper is intended to complement the Multivariable Polynomial Optimization under Constraints method and answer the question of which multivariable polynomial can be generally used in representing helicopter gas-turbine engine performance? In this sense, a variety of seven gas-turbine engines installed on different helicopters are analyzed, each one giving 512 possible polynomial models to be used for available-power calculations. While conventional statistical methods of hypothesis-testing failed in providing the answer to the question stated above of which the best general empirical model for representing engine performance is, an alternative approach based on the Singular-Value-Decomposition theorem, was proven successful in providing the answer. Moreover, this approach presented in the paper yielded a short list of 10 simple and convenient multivariable polynomials, best representing the performance of all seven engines analyzed as a group.

IPO: Interior-Point Policy Optimization under Constraints

In this paper, we study reinforcement learning (RL) algorithms to solve real-world decision problems with the objective of maximizing the long-term reward as well as satisfying cumulative constraints. We propose a novel first-order policy optimization method, Interior-point Policy Optimization (IPO), which augments the objective with logarithmic barrier functions, inspired by the interior-point method. Our proposed method is easy to implement with performance guarantees and can handle general types of cumulative multi-constraint settings. We conduct extensive evaluations to compare our approach with state-of-the-art baselines. Our algorithm outperforms the baseline algorithms, in terms of reward maximization and constraint satisfaction.

The road to Nobel prize is paved with the conceptualisations of rationality from homo economicus to homo heuristicus

In this paper we present the main psychological conceptions of rationality: unbounded rationality, bounded rationality, optimization under constraints, and ecological rationality. We show how these concepts directed the research questions, and how they shaped psychological models of complex cognitive processes. In its symbolic tradition, for more than a century, the psychology, as a fundamental cognitive science, has been focused on the question of how the environment is represented in the cognitive system, how the cognitive system operates with those information, and, ultimately, what are the outcomes of these processes. The basis on which the research efforts focusing on complex cognitive processes, such as judgment, decision-making, and reasoning - are rooted in is the stance of authors, and psychological models regarding rationality. The conceptualizations of rationality are, at the beginning of the psychological research, implicit, because they are taken from a normative approach, and the research focus is on the outcome of cognitive processes, while the functions and the processes themselves are neglected. Later, as the research diverge from the normative approach, the psychological conceptualization of rationality becomes more explicit and subjective, and more nested in the environment, and the empirical studies aim to describe the structure and dynamics of complex cognitive processes.

Helicopter gas turbine engine performance analysis: A multivariable approach

Helicopter performance relies heavily on the available output power of the engine(s) installed. A simplistic single-variable analysis approach is often used within the flight-testing community to reduce flight-test data in order to predict the available output power under various atmospheric conditions. This simplistic approach often results in unrealistic predictions. This paper proposes a novel method for analyzing flight-test data of a helicopter gas turbine engine. The so-called “Multivariable Polynomial Optimization under Constraints” method is capable of providing an improved estimation of the engine maximum available power. The Multivariable Polynomial Optimization under Constraints method relies on optimization of a multivariable polynomial model subjected to equalities and inequalities constraints. The Karush–Khun–Tucker optimization method is used with the engine operating limitations serving as inequalities constraints. The proposed Multivariable Polynomial Optimization under Constraints method is applied to a set of flight-test data of a Rolls Royce/Allison MTU250-C20 gas turbine, installed on an MBB BO-105 M helicopter. It is shown that the Multivariable Polynomial Optimization under Constraints method can predict the engine output power under a wider range of atmospheric conditions and that the standard deviation of the output power estimation error is reduced from 13 hp in the single-variable method to only 4.3 hp using the Multivariable Polynomial Optimization under Constraints method (over 300% improvement).