Magdalene Marinaki

A hybrid genetic - Particle Swarm Optimization Algorithm for the vehicle routing problem

Usually in a genetic algorithm, individual solutions do not evolve during their lifetimes: they are created, evaluated, they may be selected as parents to new solutions and they are destroyed. However, research into memetic algorithms and genetic local search has shown that performance may be improved if solutions are allowed to evolve during their own lifetimes. We propose that this solution improvement phase can be assisted by knowledge stored within the parent solutions, effectively allowing parents to teach their offspring how to improve their fitness. In this paper, the evolution of each individual of the total population, which consists of the parents and the offspring, is realized with the use of a Particle Swarm Optimizer where each of them has to improve its physical movement following the basic principles of Particle Swarm Optimization until it will obtain the requirements to be selected as a parent. Thus, the knowledge of each of the parents, especially of a very fit parent, has the possibility to be transferred to its offspring and to the offspring of the whole population, and by this way the proposed algorithm has the possibility to explore more effectively the solution space. These ideas are applied in a classic combinatorial optimization problem, the vehicle routing problem, with very good results when applied to two classic benchmark sets of instances.

A hybrid particle swarm optimization algorithm for the vehicle routing problem

This paper introduces a new hybrid algorithmic nature inspired approach based on particle swarm optimization, for successfully solving one of the most popular supply chain management problems, the vehicle routing problem. The vehicle routing problem is considered one of the most well studied problems in operations research. The proposed algorithm for the solution of the vehicle routing problem, the hybrid particle swarm optimization (HybPSO), combines a particle swarm optimization (PSO) algorithm, the multiple phase neighborhood search-greedy randomized adaptive search procedure (MPNS-GRASP) algorithm, the expanding neighborhood search (ENS) strategy and a path relinking (PR) strategy. The algorithm is suitable for solving very large-scale vehicle routing problems as well as other, more difficult combinatorial optimization problems, within short computational time. It is tested on a set of benchmark instances and produced very satisfactory results. The algorithm is ranked in the fifth place among the 39 most known and effective algorithms in the literature and in the first place among all nature inspired methods that have ever been used for this set of instances.

A Hybrid Multi-Swarm Particle Swarm Optimization algorithm for the Probabilistic Traveling Salesman Problem

The Probabilistic Traveling Salesman Problem (PTSP) is a variation of the classic Traveling Salesman Problem (TSP) and one of the most significant stochastic routing problems. In the PTSP, only a subset of potential customers need to be visited on any given instance of the problem. The number of customers to be visited each time is a random variable. In this paper, a new hybrid algorithmic nature inspired approach based on Particle Swarm Optimization (PSO), Greedy Randomized Adaptive Search Procedure (GRASP) and Expanding Neighborhood Search (ENS) Strategy is proposed for the solution of the PTSP. The proposed algorithm is tested on numerous benchmark problems from TSPLIB with very satisfactory results. Comparisons with the classic GRASP algorithm, the classic PSO and with a Tabu Search algorithm are also presented. Also, a comparison is performed with the results of a number of implementations of the Ant Colony Optimization algorithm from the literature and in 13 out of 20 cases the proposed algorithm gives a new best solution.

Honey bees mating optimization algorithm for the Euclidean traveling salesman problem

This paper introduces a new hybrid algorithmic nature inspired approach based on Honey Bees Mating Optimization for successfully solving the Euclidean Traveling Salesman Problem. The proposed algorithm for the solution of the Traveling Salesman Problem, the Honey Bees Mating Optimization (HBMOTSP), combines a Honey Bees Mating Optimization (HBMO) algorithm, the Multiple Phase Neighborhood Search-Greedy Randomized Adaptive Search Procedure (MPNS-GRASP) algorithm and the Expanding Neighborhood Search Strategy. Besides these two procedures, the proposed algorithm has, also, two additional main innovative features compared to other Honey Bees Mating Optimization algorithms concerning the crossover operator and the workers. The main contribution of this paper is that it shows that the HBMO can be used in hybrid synthesis with other metaheuristics for the solution of the TSP with remarkable results both to quality and computational efficiency. The proposed algorithm was tested on a set of 74 benchmark instances from the TSPLIB and in all but eleven instances the best known solution has been found. For the rest instances the quality of the produced solution deviates less than 0.1% from the optimum.

Particle Swarm Optimization for the Vehicle Routing Problem with Stochastic Demands

This paper introduces a new hybrid algorithmic approach based on Particle Swarm Optimization (PSO) for successfully solving one of the most popular supply chain management problems, the Vehicle Routing Problem with Stochastic Demands (VRPSD). The VRPSD is a well known NP-hard problem in which a vehicle with finite capacity leaves from the depot with full load and has to serve a set of customers whose demands are known only when the vehicle arrives to them. A number of different variants of the PSO are tested and the one that performs better is used for solving benchmark instances from the literature.

An evolutionary approach to construction of outranking models for multicriteria classification: The case of the ELECTRE TRI method

Outranking methods constitute an important class of multicriteria classification models. Often, however, their implementation is cumbersome, due to the large number of parameters that the decision maker must specify. Past studies tried to address this issue using linear and nonlinear programming, to elicit the necessary preferential information from assignment examples. In this study, an evolutionary approach, based on the differential evolution algorithm, is proposed in the context of the ELECTRE TRI method. Computational results are given to test the effectiveness of the methodology and the quality of the obtained models.

A Particle Swarm Optimization Algorithm with Path Relinking for the Location Routing Problem

This paper introduces a new hybrid algorithmic nature inspired approach based on particle swarm optimization, for solving successfully one of the most popular logistics management problems, the location routing problem (LRP). The proposed algorithm for the solution of the location routing problem, the hybrid particle swarm optimization (HybPSO-LRP), combines a particle swarm optimization (PSO) algorithm, the multiple phase neighborhood search – greedy randomized adaptive search procedure (MPNS-GRASP) algorithm, the expanding neighborhood search (ENS) strategy and a path relinking (PR) strategy. The algorithm is tested on a set of benchmark instances. The results of the algorithm are very satisfactory for these instances and for six of them a new best solution has been found.

Ant colony and particle swarm optimization for financial classification problems

Financial decisions are often based on classification models which are used to assign a set of observations into predefined groups. Such models ought to be as accurate as possible. One important step towards the development of accurate financial classification models involves the selection of the appropriate independent variables (features) which are relevant for the problem at hand. This is known as the feature selection problem in the machine learning/data mining field. In financial decisions, feature selection is often based on the subjective judgment of the experts. Nevertheless, automated feature selection algorithms could be of great help to the decision-makers providing the means to explore efficiently the solution space. This study uses two nature-inspired methods, namely ant colony optimization and particle swarm optimization, for this problem. The modelling context is developed and the performance of the methods is tested in two financial classification tasks, involving credit risk assessment and audit qualifications.

A Bilevel Genetic Algorithm for a real life location routing problem

Abstract A fast, robust and efficient algorithm for solving the Location routing problem (LRP) for one of the largest companies in Greece, which distributes wood products, is developed. To solve the companys problem, a new formulation of the LRP based on bilevel programming is proposed. Based on the fact that in the LRP, decisions are made at a strategic level and at an operational level, we formulate the problem in such a way that in the first level, the decisions of the strategic level are made, namely, the top manager finds the optimal location of the facilities: while in the second level, the operational level decisions are made, namely, the operational manager finds the optimal routing of vehicles. For this problem, a Bilevel Genetic Algorithm is proposed. The algorithm manages to provide practical solutions and significant economic benefits for the company. The algorithm is also tested on a set of benchmark instances. The results of the algorithm are also very satisfactory for those instances and, for six of them, a new best solution has been found.

Honey Bees Mating Optimization algorithm for financial classification problems

Nature inspired methods are approaches that are used in various fields and for the solution for a number of problems. This study uses a nature inspired method, namely Honey Bees Mating Optimization, that is based on the mating behaviour of honey bees for a financial classification problem. Financial decisions are often based on classification models which are used to assign a set of observations into predefined groups. One important step towards the development of accurate financial classification models involves the selection of the appropriate independent variables (features) which are relevant for the problem at hand. The proposed method uses for the feature selection step, the Honey Bees Mating Optimization algorithm while for the classification step, Nearest Neighbor based classifiers are used. The performance of the method is tested in a financial classification task involving credit risk assessment. The results of the proposed method are compared with the results of a particle swarm optimization algorithm, an ant colony optimization, a genetic algorithm and a tabu search algorithm.