A multi-objective ant colony optimization method applied to switch engine scheduling in railroad yards

Abstract

This paper proposes an ant colony optimization algorithm to assist railroad yard operational planning staff in their daily tasks. The proposed algorithm tries to minimize a multi-objective function that considers both fixed and variable transportation costs involved in moving railroad cars within the railroad yard area. This is accomplished by searching the best switch engine schedule for a given time horizon. As the algorithm was designed for real life application, the solution must be delivered in a predefined processing time and it must be in accordance with railroad yard operational policies. A railroad yard operations simulator was built to produce artificial instances in order to tune the parameters of the algorithm. The project is being developed together with industrial professionals from the Tubarão Railroad Terminal, which is the largest railroad yard in Latin America.Keywords: ACO; ant colony optimization; railroad yard operational planning; switch engine scheduling. ResumoEste trabalho propõe um algoritmo de otimização com colônia de formigas para auxiliar a equipe de planejamento de operações de pátios ferroviários em suas tarefas diárias. O algoritmo proposto é baseado em uma função multi-objetivo que busca a redução dos custos fixo e variável de movimentação de vagões no pátio. Isto é feito através da busca da melhor programação para as locomotivas de manobra, considerando um dado horizonte de planejamento. Como o algoritmo foi desenvolvido para aplicação na vida real, a solução deve ser entregue em um tempo de processamento definido previamente e deve obedecer as políticas operacionais do pátio. Foi desenvolvido um simulador de operações de pátio que gera instâncias artificiais utilizadas para ajuste dos parâmetros do algoritmo. O projeto está sendo desenvolvido em conjunto com profissionais envolvidos na operação do Terminal Ferroviário de Tubarão, o qual é o maior pátio de manobras da América Latina.

A hybrid method for a Project Scheduling Problem

“…First techniques used only constructive heuristics [3,12,15] or mathematical formulations [6,14]. However, over the last years, many metaheuristics approaches have obtained significant results in several areas [1,8,17,19,20,24,25], including validation and comparison with third-party software like [10]. They are particularly applied to the multi-mode version of RCPSP, where a task can be executed in different ways, using different amount of resources [4,5].…”

The locomotive assignment problem: a survey on optimization models

“…As reported in Sabino et al (2010), yard activities are an important part of freight transportation operations as the delays associated with these activities represent a large portion of the transit time for rail freight. Yard locomotives are often called switch engines and move cars and locomotives within the railroad yard.…”

“…Yard locomotives are often called switch engines and move cars and locomotives within the railroad yard. The solution of the LAP helps to minimize the costs of the switch operations optimizing the fleet size of the switch engines that greatly affects these costs (see Sabino et al, 2010 for more details). Lübbecke and Zimmermann (2003) report another particular real-life application of the LAP.…”

“…Computational experiments are conducted on both artificial and real-life data obtained from three different German plants (VPS, EKO, and SOL). Sabino et al (2010) propose an ant colony optimization algorithm to assist railroad yard operational planning operations. Given the information about the railroad yard layout, the switch engines available, and a detailed specification of all pending planned switch orders, the goal was to determine a switch engine schedule.…”

Methodical foundations of cargo owners’ transport services logistics management

“…Kuttner in [7] proposed an option to optimize designing networks in the cargo service of German railways, which is an interesting and original modern approach to solving the issue. Scholars J. Sabino, J. Leal, T. Stützle and M. Birattari in [8] offered an algorithm for optimizing the operational planning of the railway station performance. The proposed algorithm minimizes the multi-purpose function, which considers both fixed and variable transport costs associated with the movement of railway cars in the area of the railway station's yard.…”