Joseph M. Pasia

Solving a bi-objective vehicle routing problem by Pareto-ant colony optimization

In this paper we propose the application of Pareto ant colony optimization (PACO) in solving a bi-objective capacitated vehicle routing problem with route balancing (CVRPRB). The objectives of the problem are minimization of the tour length and balancing the routes. We propose PACO as our response to the deficiency of the Pareto-based local search (P-LS) approach, which we also developed to solve CVRPRB. The deficiency of P-LS is the lack of information flow among its pools of solutions. PACO is a natural choice in addressing this deficiency since PACO and P-LS are similar in structure. It resolves the absence of information flow through its pheromone values. Several test instances are used to demonstrate the contribution and importance of information flow among the pools of solutions. Computational results show that PACO improves P-LS in most instances with respect to different performance metrics.

Solving a bi-objective flowshop scheduling problem by pareto-ant colony optimization

In this paper we investigate the performance of pareto ant colony optimization (PACO) in solving a bi-objective permutation flowshop problem. We hybridize this technique by incorporating path relinking (PR) in four different ways. Several test instances are used to test the effectiveness of the different approaches. Computational results show that hybridizing PACO with PR improves the performance of PACO. The hybrid algorithms also show competitive results compared to other state of the art metaheuristics.

Local search guided by path relinking and heuristic bounds

In this paper we present three path relinking approaches for solving a bi-objective permutation flowshop problem. The path relinking phase is initialized by optimizing the two objectives using Ant Colony System. The initiating and guiding solutions of path relinking are randomly selected and some of the solutions along the path are intensified using local search. The three approaches differ in their strategy of defining the heuristic bounds for the local search, i.e., each approach allows its solutions to undergo local search under different conditions. These conditions are based on local nadir points. Several test instances are used to investigate the performances of the different approaches. Computational results show that the decision which allows solutions to undergo local search has an influence in the performance of path relinking. We also demonstrate that path relinking generates competitive results compared to the best known solutions of the test instances.

A population-based local search for solving a bi-objective vehicle routing problem

In this paper we present a population-based local search for solving a bi-objective vehicle routing problem. The objectives of the problem are minimization of the tour length and balancing the routes. The algorithm repeatedly generates a pool of good initial solutions by using a randomized savings algorithm followed by local search. The local search uses three neighborhood structures and evaluates the fitness of candidate solutions using dominance relation. Several test instances are used to assess the performance of the new approach. Computational results show that the population-based local search outperforms the best known algorithm for this problem.

Path relinking on many-objective NK-landscapes

Path relinking is a population-based heuristic that explores the trajectories in decision space between two elite solutions. It has been successfully used as a key component of several multi-objective optimizers, especially for solving bi-objective problems. In this paper, we focus on the behavior of pure path relinking, propose several variants of the path relinking that vary on their selection strategies, and analyze its performance using several many-objective NK-landscapes as instances. The study shows that the path relinking becomes more effective in improving the convergence of the algorithm as the number of objectives increases. It also shows that the selection strategy associated to path relinking plays an important role to emphasize either convergence or spread of the algorithm.

Improving multi-objective random one-bit climbers on MNK-landscapes

Multi-objective random one-bit climbers (moRBCs) are one class of stochastic local search-based algorithms that maintain a reference population of solutions to guide their search. They have been shown to perform well in solving multi-objective optimization problems. In this work, we further enhance the moRBCs by introducing tabu moves to improve their efficiency and search for more promising solutions. We also improve the selection to update the reference population and archive using a procedure that provides better mechanism to preserve diversity among the solutions. We use several MNK-landscape models to study the behavior of the modified moRBCs.