Alberto García-Villoria

A MILP model to design hybrid wind–photovoltaic isolated rural electrification projects in developing countries

Electrification systems based on the use of renewable energy sources are a suitable option for providing electricity to isolated communities autonomously. Wind and hybrid wind–photovoltaic (PV) systems are increasingly getting attention. To electrify scattered communities, designs that combine individual systems and microgrids have recently proven advantageous. In this paper we present a mathematical programming model to optimize the design of hybrid wind–PV systems that solves the location of the wind–PV generators and the design of the microgrids, taking into account the demand of the consumption points and the energy potential. The criterion is the minimization of the initial investment cost required to meet the demand. The proposed hybrid model is tested with realistic size instances and results show the instances are efficiently solved. Moreover, the model is applied to real case studies in Peru; obtained results verify that the hybrid model efficiently finds solutions that significantly reduce costs.

Solving the response time variability problem by means of the electromagnetism-like mechanism

The response time variability problem (RTVP) is an NP-hard combinatorial scheduling problem that has been recently formalised in the literature. The RTVP has a wide range of real-life applications such as in the automobile industry, when models to be produced on a mixed-model assembly line have to be sequenced under a just-in-time production. The RTVP occurs whenever products, clients or jobs need to be sequenced so as to minimise variability in the time between the instants at which they receive the necessary resources. In two previous studies, three metaheuristic algorithms (a multi-start, a GRASP and a PSO algorithm) were proposed to solve the RTVP. We propose solving the RTVP by means of the electromagnetism-like mechanism (EM) metaheuristic algorithm. The EM algorithm is based on an analogy with the attraction-repulsion mechanism of the electromagnetism theory, where solutions are moved according to their associated charges. In this paper we compare the proposed EM metaheuristic procedure with the three metaheuristic algorithms aforementioned and it is shown that, on average, the EM procedure improves strongly on the obtained results.

Hyper-heuristic approaches for the response time variability problem

We propose two classes for the implementation of hyper-heuristic algorithms. The first is based on constructive heuristics, whereas the second uses improvement methods. Within the latter class, a general framework is designed for the use of local search procedures and metaheuristics as low-level heuristics. A dynamic scheme to guide the use of these approaches is also devised. These ideas are tested on an NP-hard scheduling problem known as the response time variability problem (RTVP). An intensive computational experiment shows, especially in the second class where the new best results are found, the effectiveness of the proposed hyper-heuristics.

Solving the response time variability problem by means of a psychoclonal approach

The response time variability problem (RTVP) is a combinatorial scheduling problem that has recently appeared in the literature. This problem has a wide range of real life applications in fields such as manufacturing, hard real-time systems, operating systems and network environments. Originally, the RTVP occurs whenever products, clients or jobs need to be sequenced in such a way that the variability in the time between the instants at which they receive the necessary resources is minimized. Since the RTVP is hard to solve, heuristic techniques are needed for solving it. Three metaheuristic—multi-start, GRASP and PSO—algorithms proposed for solving the RTVP in two previous studies have been the most efficient to date in solving non-small instances of the RTVP. We propose solving the RTVP by means of a psychoclonal algorithm based approach. The psychoclonal algorithm inherits its attributes from Maslow’s hierarchy of needs theory and the artificial immune system (AIS) approach, specifically the clonal selection principle. In this paper, we compare the proposed psychoclonal algorithm with the three aforementioned metaheuristic algorithms and show that, on average, the psychoclonal algorithm strongly improves on the results obtained.

Solving the response time variability problem by means of the cross-entropy method

The response time variability problem (RTVP) is an NP-hard combinatorial scheduling problem that has recently appeared in the literature. The RTVP has a wide range of production line systems applications such as sequencing the models to be produced on a mixed-model assembly line in a just-in-time context. This problem occurs whenever several units of different models need to be sequenced so as to minimise the variability of the distance between any two consecutive units of the same model. A mathematical mixed integer linear programming (MILP) model has been presented by another study, but the practical limit for obtaining optimal solutions is around 40 units to be scheduled. Another study has developed five heuristic algorithms to solve non-small RTVP instances. We propose to solve the RTVP by means of the metaheuristic cross-entropy (CE) method, which has been developed recently. We report on the computational experiments in which the CE method is compared with the five heuristic algorithms proposed in the literature.

Scheduling commercial advertisements for television

The problem of scheduling the commercial advertisements in the television industry is investigated. Each advertiser client demands that the multiple airings of the same brand advertisement should be as spaced as possible over a given time period. Moreover, audience rating requests have to be taken into account in the scheduling. This is the first time this hard decision problem is dealt with in the literature. We design two mixed integer linear programming (MILP) models. Two constructive heuristics, local search procedures and simulated annealing (SA) approaches are also proposed. Extensive computational experiments, using several instances of various sizes, are performed. The results show that the proposed MILP model which represents the problem as a network flow obtains a larger number of optimal solutions and the best non-exact procedure is one that uses SA.

Combining matheuristics and MILP to solve the accessibility windows assembly line balancing problem level 2 (AWALBP-L2)

Abstract We propose an approach combining a matheuristic and a MILP model to solve the variant Level 2 of the Accessibility Windows Assembly Line Balancing Problem (AWALBP-L2). This is a novel problem that arises in those real-world assembly lines where, in contrast to the most common ones, the length of the workpieces is larger than the widths of the workstations. This means that, at any time, a workstation cannot access an entire workpiece, but only a restricted portion of a workpiece or two consecutive workpieces. As a result, a workstation can only perform, at any time, the subset of tasks that fall inside its accessible area. The problem is to solve the task assignment and the movement scheme subproblems, while minimizing the cycle time. The proposed solving approach consists of (i) a matheuristic to generate good feasible solutions and compute bounds and (ii) a MILP model that makes use of the obtained bounds. A computational study is carried out to compare the performance of the proposed approach with the existing literature.

A MILP model for the Accessibility Windows Assembly Line Balancing Problem (AWALBP)

This work studies a novel assembly line balancing problem that has recently appeared in the literature, which we name Accessibility Windows Assembly Line Balancing Problem (AWALBP). AWALBP is a real-world industrial problem that arises in those assembly lines where, as opposed to the most common ones, the length of the workpiece is larger than the widths of the workstation. This means that, at any time, a workstation cannot access one whole workpiece, but only a restricted portion of one or two consecutive workpiece. In our problem the cycle decomposes into stationary stages separated by forward steps, according to a cyclic movement scheme. The aim of this paper is (i) to formalise the AWALBP and its variants, and (ii) to propose a Mixed Integer Linear Programming (MILP) model using two alternative formulations to solve the variant AWALBP-L2. This variant involves solving the task assignment and the movement scheme sub-problems (with the objective of minimising the cycle time). An extensive computational experiment is carried out to study the behaviour of the proposed model for different instance sizes. To the best of our knowledge, this is the first work in the literature which provides optimal solutions for AWALBP-L2. In addition, a set of benchmark instances is provided, which can be further used by the research community.

Heuristics and simulated annealing procedures for the accessibility windows assembly line problem level 1 (AWALBP-L1)

Abstract In some assembly lines, the workpieces are larger than the workstations. This implies that at a given instant the workstations have access to only a portion of the workpieces. In this context, the accessibility windows assembly line balancing problem (AWALBP) arises. In the AWALBP, the cycle is split into forward steps and stationary stages. The workpieces advance during the forward steps and the tasks are processed during the stationary stages. In each stationary stage, the workstations have access to different parts of the workpieces. This work solves the first level of AWALBP (AWALBP-L1), which consists in assigning the tasks among the workstations and stationary stages. Specifically, it is considered the AWALBP-L1 case in which the tasks can be processed in several workstations and their processing times depend on the workstation in which the tasks are processed. To solve the problem, we propose several heuristics and simulated annealing procedures. An extensive computational experiment is carried out to evaluate their performance.

A heuristic procedure for solving the Lexicographic Bottleneck Assembly Line Balancing Problem (LB-ALBP)

The Lexicographic Bottleneck Assembly Line Balancing Problem (LB-ALBP) is a new assembly-line balancing problem recently defined in the literature. The LB-ALBP hierarchically minimises the workload of the most heavily loaded workstation, followed by the workload of the second most heavily loaded workstation, followed by the workload of the third most heavily loaded workstation, and so on. The original study presents two mixed-integer linear programming (MILP) models designed to solve the LB-ALBP optimally, together with three heuristic procedures based on these MILPs. In this paper, we propose and test new algorithms that combine a heuristic procedure for obtaining an initial solution and several local search procedures, which are an improvement upon the heuristic procedures published to date.