Chair: Maria Antónia Carravilla, INESC TEC, Faculdade de Engenharia, Universidade do Porto, Portugal
An Agent-based Approach to Schedule Crane Operations in Rail-Rail Transshipment Terminals
Sam Heshmati (firstname.lastname@example.org), Zafeiris Kokkinogenis (email@example.com), Rosaldo Rossetti (firstname.lastname@example.org), Maria Antónia Carravilla (email@example.com), José Fernando Oliveira (firstname.lastname@example.org)
Rail-rail transshipment terminals (RRTT) play the hub role in hub-and-spoke rail networks. In this type of terminals, containers are transshipped among freight trains by gantry cranes. This study aims scheduling crane operations with an agent-based simulation approach. The decisions to make include: the positions of containers on outbound trains, the assignment of container moves to cranes, as overlapping areas for the cranes’ operation are considered, the sequence of transshipments. The goal is to minimize the total transshipment time. The contribution of this paper is an agent-based approach to the crane operations scheduling problem in a RRTT framework. This approach is validated on a set of instances taken from the literature and the results compared against a variable neighbourhood descent algorithm.
Keywords: Railway systems, Agent-based simulation, Scheduling crane operations
A MIP Model for Production Planning in the Roasting Coffee Industry
Diana Ospina (email@example.com), Maria Antónia Carravilla (firstname.lastname@example.org), José Fernando Oliveira (email@example.com)
The coffee supply chain includes harvesting, comercialization, production and distribution. This paper presents a case study of a Portuguese roasted coffee company. The production process in the company begins with the storage in warehouses and in silos and continues through blending, roasting, grinding and finally packaging and warehousing. These processes are carried out in order to fulfil different requirements in terms of freshness, aroma, flavour and colour of the coffee. A mixed integer (MIP) model for production planning has been built, taking into account as main decision variables, the type of coffee to load in each silo in each period. The objective is to produce the coffee as near as possible to a due date in order to maintain the freshness and to satisfy the demand of the clients. The company needs the model to embed in a decision support system that allows them to test the acceptance of coffee orders and to define at the beginning of each day how the silos must be loaded. The results are very encouraging because it is possible to test many scenarios for the orders in a short period of time.
Keywords: Coffee production, Coffee supply chain, Production planning
An algorithm for solving real-world routing and scheduling problems
Nico Kyngäs (firstname.lastname@example.org), Kimmo Nurmi (email@example.com), Jari Kyngäs (firstname.lastname@example.org)
Combined routing and scheduling problems are extremely challenging combinatorial optimization problems that have plenty of real-world applications. In a basic problem setting a host of mobile resources (vehicles, employees etc.) must be used to carry out tasks (cleaning, home care, guarding, repairing, delivery of goods, newspaper etc.) at spatially diverse locations. A schedule or route of a vehicle then consists of a sequence of tasks and arrival times to their locations. Usually the goal is to minimize the cost of the operation under a set of constraints. The most important source of cost is employees. The other one is the total length of the routes, which affects fuel consumption, the total working hours of the employees, the wear and tear of the vehicles and often the number of vehicles and/or employees required. Constraints include shift rules laid down in the collective labour agreement (maximum shift length, placement of breaks etc.) and requirements of the tasks (time windows, skills, equipment etc.). Other, often relevant yet immeasurable considerations include environmental concerns and customer and employee satisfaction. The PEAST algorithm we have developed is a population-based local search method that uses several metaheuristic concepts to avoid getting stuck on local optima. It has been used for solving various combinatorial real-world scheduling problems, such as workforce scheduling, sports scheduling and school timetabling. While we have solved numerous workforce scheduling problems for our business partners, we have been approached regarding solutions for routing and scheduling problems. Therefore we adapted the PEAST algorithm for this type of problems. We demonstrate its viability on the well-known theoretical benchmark instances, while our ultimate goal is to solve complex real-world instances of our business partners.
Keywords: Workforce scheduling, routing, PEAST algorithm