Marcello Sammarra

The Berth Allocation Problem: A Strong Formulation Solved by a Lagrangean Approach

The berth allocation problem is one of the most relevant logistics problems arising in the management of container ports. Depending on assumptions made on the berthing location policy, two classes of the berth allocation problem have been considered in literature: the discrete case and the continuous case. In this paper, the properties of the discrete berth allocation problem, formulated as a dynamic scheduling problem, are analyzed. A new formulation of the problem is proposed, which is shown to be more compact and stronger than another one from the literature; a Lagrangean heuristic algorithm is developed; and computational results are presented.

A column generation heuristic for a dynamic generalized assignment problem

This paper studies the dynamic generalized assignment problem (DGAP) which extends the well-known generalized assignment problem by considering a discretized time horizon and by associating a starting time and a finishing time with each task. Additional constraints related to warehouse and yard management applications are also considered. Three linear integer programming formulations of the problem are introduced. The strongest one models the problem as an origin-destination integer multi-commodity flow problem with side constraints. This model can be solved quickly for instances of small to moderate size. However, because of its computer memory requirements, it becomes impractical for larger instances. Hence, a column generation algorithm is used to compute lower bounds by solving the linear program (LP) relaxation of the problem. This column generation algorithm is also embedded in a heuristic aimed at finding feasible integer solutions. Computational experiments on large-scale instances show the effectiveness of the proposed approach.

Quay crane scheduling with time windows, one-way and spatial constraints

In this paper, we address the quay crane scheduling problem in maritime container terminals, with some new features on the machine environment. We consider rail-mounted cranes and we assume that they are not allowed to span all over the vessel and that they can work only in predefined time windows. We furthermore take into account one-way constraints on the crane movements. We present a mixed integer mathematical model and a heuristic algorithm to get feasible solutions to the problem. Finally, we report on some computational experience.

The Terminal-Oriented Ship Stowage Planning Problem

The Ship Stowage Planning Problem is the problem of determining the optimal position of containers to be stowed in a containership. In this paper we address the problem considering the objectives of the terminal management that are mainly related to the yard and transport operations. We propose a Binary Integer Program and a two-step heuristic algorithm. An extensive computational experience shows the efficiency and effectiveness of our approach. A classification scheme for stowage planning problems is also provided.

Agent-based simulation for the evaluation of a new dispatching model for the straddle carrier pooling problem

To deal with several classes of planning and management problems in different business domains (transport, energy, defense, etc.) and at different decisional levels (strategic, tactical and operational levels), significant benefits can derive from the definition of solution methods which are able to combine and capitalize on the strengths of mathematical programming and agent-based techniques. In this paper, an approach based on the joint use of mathematical programming and agent-based techniques is exploited in the context of maritime logistics. In particular, a transshipment terminal is considered with the goal of planning the container transportation between the yard and the quay during the loading/discharging operations of a containership by exploiting a pool of straddle carriers shared among the quay cranes. The goal is to route the straddle carriers by determining a sequence of container movements for each of them in order to optimize a given performance criterion. In particular, a new dispatching model is proposed and evaluated through an agent-based simulation model which also made it possible a better understanding of interesting properties of the problem under consideration.

Some observations about the extreme points of the Generalized Cardinality-Constrained Shortest Path Problem polytope

The Generalized Cardinality-Constrained Shortest Path Problem (GCCSPP) consists in finding the minimum cost path in a digraph, using at most r arcs in a subset F of the arc set. We propose an algebraic characterization of the extreme points of the associated polytope, and then we show that it is equivalent to the geometric one, obtained extending to the GCCSPP some known results for the cardinality-constrained shortest path problem.

Skipping the Storage Phase in Container Transshipment Operations

The paper deals with the problem of scheduling the loading/discharging operations of two simultaneously berthed vessels, assuming that some of the containers discharged from a vessel must be directly loaded on the other one. For these containers also the stowage position must be decided. The aim is to minimize the time needed to complete all the operations required by the involved vessels. For this problem we present a mathematical model, a heuristic algorithm and discuss the computational results on a set of randomly generated instances.