Alberto Gomez

Forecasting next-day price of electricity in the Spanish energy market using artificial neural networks

In this paper, next-day hourly forecasts are calculated for the energy price in the electricity production market of Spain. The methodology used to achieve these forecasts is based on artificial neural networks, which have been used successfully in recent years in many forecasting applications. The days to be forecast include working days as well as weekends and holidays, due to the fact that energy price has different behaviours depending on the kind of day to be forecast. Besides, energy price time series are usually composed of too many data, which could be a problem if we are looking for a short period of time to reach an adequate forecast. In this paper, a training method for artificial neural nets is proposed, which is based on making a previous selection for the multilayer perceptron (MLP) training samples, using an ART-type neural network. The MLP is then trained and finally used to calculate forecasts. These forecasts are compared to those obtained from the well-known Box-Jenkins ARIMA forecasting method. Results show that neural nets perform better than ARIMA models, especially for weekends and holidays. Both methodologies calculate more accurate forecasts-in terms of mean absolute percentage error-for working days that for weekends and holidays. Agents involved in the electricity production market, who may need fast forecasts for the price of electricity, would benefit from the results of this study.

A knowledge-based evolutionary strategy for scheduling problems with bottlenecks

Abstract In this paper we confront a family of scheduling problems by means of genetic algorithms: the job shop scheduling problem with bottlenecks. Our main contribution is a strategy to introduce specific knowledge into the initial population. This strategy exploits a probabilistic-based heuristic method that was designed to guide a conventional backtracking search. We report experimental results on two benchmarks, the first one includes a set of small problems and is taken from the literature. The second includes medium and large size problems and is proposed by our own. The experimental results show that the performance of the genetic algorithm clearly augments when the initial population is seeded with heuristic chromosomes, the improvement being more and more appreciable as long as the size of the problem instance increases. Moreover premature convergence which sometimes appears when randomness is limited in any way in a genetic algorithm is not observed.

A review of machine learning in dynamic scheduling of flexible manufacturing systems

A common way of dynamically scheduling jobs in a flexible manufacturing system (FMS) is by means of dispatching rules. The problem of this method is that the performance of these rules depends on the state the system is in at each moment, and no single rule exists that is better than the rest in all the possible states that the system may be in. It would therefore be interesting to use the most appropriate dispatching rule at each moment. To achieve this goal, a scheduling approach which uses machine learning can be used. Analyzing the previous performance of the system (training examples) by means of this technique, knowledge is obtained that can be used to decide which is the most appropriate dispatching rule at each moment in time. In this paper, a review of the main machine learning-based scheduling approaches described in the literature is presented.

RESOLUTION OF STRIP-PACKING PROBLEMS WITH GENETIC ALGORITHMS

This paper studies strip-packing problems. It is our aim to optimise the position of a number of rectangular shapes on a base surface in order to minimise wastage of material. As the problem is a complex NP-complete one, a heuristic based on genetic algorithms (GA) is used to solve it. The main problem is the wide variety of genetic algorithms available in the literature, which makes it hard to know which variation is best suited to this type of problem. We conclude that using a cyclic crossover GA with fitness by area and variable mutation works best for this problem.

A Decision-Making Support System on a Products Recovery Management Framework. A Fuzzy Approach:

The considerable amount of uncertainty involved in defining the factors that affect reverse logistics (RL) decision-making and the complex interrelationships between those factors make it rather difficult to decide what recovery policy a business should pursue. This article proposes a fuzzy system that helps in such decision-making and thereby mitigates these difficulties. The knowledge related to the decision is incorporated into the system by means of conditional rules, which serve to provide the ideal recovery policy for each particular case. The model proposed is applied to the analysis of a number of examples and proves to be a versatile tool that provides coherent results. These characteristics could be of critical importance especially in the point of entry into the RL pipeline and in the centralized return centres.

Review of the current state and main sources of dioxins around the world

Polychlorinated dibenzo-p-dioxins (PCDDs) and polychlorinated dibenzofurans (PCDFs) are a group of dangerous compounds, emitted mostly from anthropogenic sources, that have negative effects on human health. Therefore, it is interesting to analyze the emission patterns of dioxins proceeding from different sources around the world, to observe the actual trend of the transmission of dioxins and furans into the atmosphere.For that reason, the main objective of the present document is to provide a general assessment about the dioxin problematic, analyzing the main parameters that influence the ambient concentration of dioxins worldwide, and describing the most characteristic features of the fingerprint from different sources, while making emphasis in the importance that non-industrial sources are gaining over the last years in front of the decreasing tendency of industrial sources. The description of the most important abatement technologies for dioxins is also included in this review. Implications: Given the negative effects of dioxins in human health, it is important to depict and locate the main sources of these dangerous compounds. Emissions proceeding from industrial facilities have decreased over the last years; however, other zones where nonindustrial sources used to be relevant contributors do not show the same decreasing tendency because it is more difficult to control this type of emissions. For that reason, future studies should focus on measuring and regulating this highly uncontrolled source of dioxins.

Genetic algorithms to optimise the time to make stock market investment

The application of Artificial Intelligence described in this article is intended to resolve the issue of speculation on the stock market. Genetic Algorithms is the technique that is used, with the article focusing on the different ways that chromosomes can be designed and on how the pertinent evaluation mechanism is established. The problem will be based on the speculation systems that are typical of Technical Analysis.

Solving capacitated vehicle routing problem by artificial bee colony algorithm

This paper presents a new Artificial Bee Colony algorithm for solving the capacitated vehicle routing problem. The main novel characteristic of the proposed approach relies upon an efficient way of coordinating, for each group of bees, a well-defined focus of work. In the algorithm, we provide two specializations namely diversification and intensification where the former is controlled by the employed and the scout bees whereas the latter by the onlookers. The two datasets commonly used as benchmark instances are used to assess the performance of the proposed algorithm. The results show that the proposed algorithm obtains interesting results.

DYNAMIC SCHEDULING OF MANUFACTURING SYSTEMS WITH MACHINE LEARNING

A common way of scheduling jobs dynamically in a manufacturing system is by means of dispatching rules. The drawback of this method is that the performance of these rules depends on the state the system is in at each moment, and no one rule exists that overrules the rest in all the possible states that the system may be in. It would therefore be interesting to use the most appropriate rule at each moment. To achieve this goal, a scheduling approach which uses machine learning is presented in this paper. The methodology proposed in this paper may be divided into five basic steps. Firstly, definition of the appropriate control attributes for identifying the relevant manufacturing patterns. In second place, creation of a set of training examples using different values of the control attributes. Subsequently, acquiring of heuristic rules by means of a machine learning program. Then, using of the previously calculated heuristic rules to select the most appropriate dispatching rules, and finally testing of the performance of the approach. The approach that we propose is applied to a flow shop system and to a classic job shop configuration. The results demonstrate that this approach produces an improvement in the performance of the system when compared to the traditional method of using dispatching rules.

Dynamic scheduling of manufacturing systems using machine learning: An updated review

Abstract A common way of dynamically scheduling jobs in a manufacturing system is by implementing dispatching rules. The issues with this method are that the performance of these rules depends on the state the system is in at each moment and also that no “ideal” single rule exists for all the possible states that the system may be in. Therefore, it would be interesting to use the most appropriate dispatching rule for each instance. To achieve this goal, a scheduling approach that uses machine learning can be used. Analyzing the previous performance of the system (training examples) by means of this technique, knowledge is obtained that can be used to decide which is the most appropriate dispatching rule at each moment in time. In this paper, a literature review of the main machine learning based scheduling approaches from the last decade is presented.