Abstract
Optimization in current decision support systems has a highly interdisciplinary nature related with the need to integrate different techniques and paradigms for solving real-world complex problems. Computing optimal solutions in many of these problems are unmanageable. Heuristic search methods are known to obtain good results in an acceptable time interval. However, parameters need to be adjusted to allow good results. In this sense, learning strategies can enhance the performance of a system, providing it with the ability to learn, for instance, the most suitable optimization technique for solving a particular class of problems, or the most suitable parameterization of a given algorithm on a given scenario. Hyper-heuristics arise in this context as efficient methodologies for selecting or generating (meta) heuristics to solve NP-hard optimization problems. This paper presents the specification of a hyper-heuristic for selecting techniques inspired in nature, for solving the problem of scheduling in manufacturing systems, based on previous experience. The proposed hyper-heuristic module uses a reinforcement learning algorithm, which enables the system with the ability to autonomously select the meta-heuristic to use in optimization process as well as the respective parameters. A computational study was carried out to evaluate the influence of the hyper-heuristics on the performance of a scheduling system. The obtained results allow to conclude about the effectiveness of the proposed approach.

Abstract

Abstract
Schedulability analyses, while valuable in theoretical research, cannot be used in practice to reason about the timing behaviour of a real-time system without including the overheads induced by the implementation of the scheduling algorithm. In this paper, we provide an overhead-aware schedulability analysis based on demand bound functions for two hard real-time semi-partitioned scheduling algorithms, EDFWM- and C=D. This analysis is based on a novel implementation that uses a global clock to reduce the overheads incurred due to the release jitter of migrating subtasks. The analysis is used to guide the respective off-line task assignment and splitting procedures. Finally, results of an evaluation are provided highlighting how the different algorithms perform with and without a consideration of overheads.

Abstract
Net Present Value (NPV), Weighted Average Cost of Capital (WACC), Internal Rate of Return (IRR), and Total-Life Cost of Capital (TLCC) are economic concepts widely used in capital budgeting to measure and compare the profitability of investments. More specifically, in the electricity sector these measures, with the Levelized Cost of Energy (LCOE), are very often used to assess investments in generation assets. At the same time, electricity generation models based on mathematical programming and game theory have also been developed to determine optimal expansion plans of the generation capacity for long-term horizons. Though these two techniques have both been applied in the literature to assess generation investments in the electricity sector, taking into account, among others, investment, maintenance and operation costs, their mathematical relationships have been rarely reported or even understood. Here we provide some insight on the mathematical links existing between both approaches. © 2015 IEEE.

Abstract
Optimal operation of power systems with high integration of renewable power sources has become difficult as a consequence of the random nature of some sources like wind energy and photovoltaic energy. Nowadays, this problem is solved using Monte Carlo Simulation (MCS) approach, which allows considering important statistical characteristics of wind and solar power production such as the correlation between consecutive observations, the diurnal profile of the forecasted power production, and the forecasting error. However, MCS method requires the analysis of a representative amount of trials, which is an intensive calculation task that increases considerably with the number of scenarios considered. In this paper, a model to the scheduling of power systems with significant renewable powergeneration based on scenario generation/reduction method, which establishes a proportional relationship between the number of scenarios and the computational time required to analyse them, is proposed. The methodology takes information from the analysis of each scenario separately to determine the probabilistic behaviour of each generator at each hour in the scheduling problem. Then, considering a determined significance level, the units to be committed are selected and the load dispatch is determined. The proposed technique was illustrated through a case study and the comparison with stochastic programming approach was carried out, concluding that the proposed methodology can provide an acceptable solution in a reduced computational time.

Abstract
In this paper, we characterize two power indices introduced in [1] using two different modifications of the monotonicity property first stated by [2]. The sets of properties are easily comparable among them and with previous characterizations of other power indices.