João Miguel Lemos Chasqueira Nabais

A novel predictive control based framework for optimizing intermodal container terminal operations

Due to the increase in world-wide containerized cargo transport port authorities are facing considerable pressure to increase efficiency of existing facilities. Container vessels with 18,000 TEUs (twenty-foot equivalent units) are expected soon to create high flow peaks at container terminals. In this paper we propose a new framework for managing intermodal container terminals, based on the model predictive control methodology. A model based on queues and container categorization is used by a model predictive controller to solve the handling resource allocation problem in a container terminal in an optimal way, while respecting constraints on resource availability. The optimization of the operations is performed in an integrated way for the whole terminal rather than only for an individual subprocess. Containers are categorized into empty and full containers, and divided in classes according to their final destination. With more detailed information available, like container final destination, it is possible to establish priorities for the container flows inside the terminal. The order in which the container classes should be loaded into a carrier can now be addressed taking into account the carrier future route. The model ability to track the number of containers per class makes this framework suitable for describing terminals integrated in an intermodal transport network and a valuable tool for coordinating the transport modal shift towards a more sustainable and reliable transport. The potential of the proposed framework is illustrated with simulation studies based on a high-peak flow scenario and for a long-term scheduled scenario.

Model predictive control for a sustainable transport modal split at intermodal container hubs

The increase of international commerce and the expected container vessels capacity with 18, 000 TEU (twenty-foot equivalent unit) will put a considerable pressure on container hubs. High flow peaks will appear at gateway hubs in the transport network compromising the cargo transportation towards the hinterland and decreasing the network transport capacity. Moreover, authorities are forcing transport operators to operate in more sustainable ways. For container hubs this is translated into making a preferable choice for barge and train modalities before opting for truck modality. In this work we present a framework based on Model Predictive Control (MPC) to address the so-called transport modal split problem for the outgoing cargo at container hubs. We use two features (destination and due time) to categorize the cargo present at a container hub and develop a dynamic model to make predictions of cargo volume over time. The controller decision takes into account transporting cargo towards the final destination while opting for sustainable transport modalities. The approach is able to assign cargo in advance to the existing connections at the hub in order to overcome predicted cargo peaks in the future. The framework can also be used to choose between different connection schedules. Giving decision freedom to container hubs is a step towards a synchromodal and more flexible transport network. These statements are illustrated with two simulation examples.

Setting cooperative relations among terminals at seaports using a multi-agent system

Seaports are gateways between the over sea and the hinterland commerce, where different cargo types are handled at dedicated terminals. Currently, seaports are facing traffic congestion leading to a decrease in its performance. Prior to increase the existing infrastructures in terms of transport capacity between the seaport and the hinterland it is important to improve cooperation among terminals. A multi-agent system to guarantee cooperation among terminals within a seaport is proposed in this paper. A control agent is assigned to each terminal and is responsible for the cargo assignment to the transport capacity at its disposal such that cargo arrives on time at the agreed location. Control agents solve in parallel an optimization problem formulated in accordance to the Model Predictive Control (MPC) strategy. Cooperation among control agents is established using a coordinator agent that updates the transport capacity assigned to each control agent based on the marginal costs provided by all control agents. The proposed framework does not require the exchange of private information and assumes an altruist behavior for all control agents. The proposed approach can perform similarly to a central approach. The framework performance is illustrated with simulation studies considering a seaport composed of 3 container terminals.

Hierarchical Model Predictive Control for optimizing intermodal container terminal operations

Transportation networks are large-scale complex spatially distributed systems whose purpose is to deliver commodities at the agreed time and at the agreed location. The network nodes (terminals, depots or warehouses) can be seen as the main decision making centers, as there the different economic actors interact with each other. In particular, the intermodal container terminal is responsible for storing containers until they are picked up for transport towards their final destination. Operations management at intermodal container terminals can be seen as a flow assignment problem. In this work we present a Hierarchical Model Predictive Control (HMPC) framework for addressing flow assignments in intermodal container terminals. The approach proposed is original due to its capability to keep track of the container class while solving a flow assignment problem respecting the available resources. However, the dimension of the problem to be solved grows with the number of container classes handled and the number of available connections at the terminal. A system decomposition inspired by container flows related to each connection served at the terminal is proposed to diminish the problem dimension to solve. The framework proposed is easily scalable to container terminals where hundreds of container classes and connections are available. The potential of the proposed framework is compared to a centralized Model Predictive Control (MPC) framework and is illustrated with a simulation study based on a long-term scheduled scenario.

A constrained MPC heuristic to achieve a desired transport modal split at intermodal hubs

Intermodal hubs are a component of freight transportation networks that have as main goal to deliver cargo at the agreed time and at the agreed location. Currently, authorities are forcing transport operators to act in more sustainable ways. For intermodal hubs this is translated into making a preferable choice for sustainable transport modalities. In some cases, this is no longer a choice and is imposed on the intermodal hub in terms of a desired transport modal split. In this paper, a heuristic based on Model Predictive Control (MPC) to achieve a desired transport modal split at intermodal hubs is proposed. A terminal state constraint is used for the quantity of cargo assigned per modality over the prediction horizon to guide the cargo assignment. Feasibility of the optimization problem and cargo delivery at the agreed time are assured by relaxing the terminal state constraint. The proposed heuristic can anticipate the transport of cargo due to the inclusion of predictions, leading to a push of cargo towards the final destination. As cargo is moving in anticipation to the due time the transport is more robust to unforseen events, such as jams and weather conditions. The proposed heuristic is a step towards sustainable and synchromodal transportation networks. Simulation experiments illustrate the validity of these statements.

Hierarchical MPC for Multiple Commodity Transportation Networks

Transportation networks are large scale complex systems spatially distributed whose objective is to deliver commodities at the agreed time and at the agreed location. These networks appear in different domain fields, such as communication, water distribution, traffic, logistics and transportation. A transportation network has at the macroscopic level storage capability (located in the nodes) and transport delay (along each connection) as main features. Operations management at transportation networks can be seen as a flow assignment problem. The problem dimension to solve grows exponentially with the number of existing commodities, nodes and connections. In this work we present a Hierarchical Model Predictive Control (H-MPC) architecture to determine flow assignments in transportation networks, while minimizing exogenous inputs effects. This approach has the capacity to keep track of commodity types while solving the flow assignment problem. A flow decomposition of the main system into subsystems is proposed to diminish the problem dimension to solve in each time step. Each subsystem is managed by a control agent. Control agents solve their problems in a hierarchical way, using a so-called push-pull flow perspective. Further problem dimension reduction is achieved using contracted projection sets. The framework proposed can be easily scaled to network topologies in which hundreds of commodities and connections are present.

New Fault Isolation Architecture for Irrigation Canals

Abstract Irrigation is one of the most consuming water resources in human activity. As water conveyance systems are commonly spatially distributed and crossing large regions it is important to guarantee their best efficiency. A fault diagnosis architecture is an important tool to increase efficiency in water conveyance systems. In the presence of leaks, unauthorized water extractions or water level sensor faults, the service level can be severely compromised. Recent literature deal with these situations as a unique fault of water extraction type. Isolating correctly each fault is important to access the real current state of the irrigation canals and proceed in accordance to restore its nominal conditions. This paper proposes a fault diagnosis architecture to distinguish common faults in water canals, considering either unexpected water extractions, gate faults and downstream water level sensor errors. The architecture is based on geometric and hydraulic parameters and therefore can be extended to existent irrigation canals. The proposed fault isolation architecture is an important tool to support maintenance services and be extended to fault tolerant controllers.

Flexible Discrete Time State Space Model for Canal Pools

Water is a vital resource for mankind used in activities such as agriculture, industry and domestic activity. Irrigation is one of the most consuming water resources in human activity. Irrigation canals are characterized for being spatially distributed crossing different administrative regions. As water is becoming a scarce and valuable resource, efficient engineering water conveyance networks are required. In this paper a discrete state space for modeling open-channels is presented. The well known Saint-Venant equations are first linearized for a steady state and then discretized using the Preissmann scheme. The resulting model is shown to be computational simple and flexible to accommodate different type of boundary conditions, in flow, water depth or hydraulic structures dynamics, which are important features for modeling complex water conveyance systems. The hydraulic model also offers monitoring ability along the canal axis and can therefore be integrated in fault diagnosis and tolerant control strategies. The model is validated with experimental data from a real canal property of the Evora University.

Data Based Modeling of a Large Scale Water Delivery System

Water is a vital resource and the growing populations and economies around the globe are pushing its demand worldwide. Therefore, the water conveyance operation should be well managed and improved. This paper proposes the development of reliable models able to predict water levels of a real 24.4 km water delivery channel in real time. This is a difficult task because this is a time-delayed dynamical system distributed over a long distance with nonlinear characteristics and external perturbations. Artificial neural networks are used, which are a well-known modeling technique that has been applied to complex and nonlinear systems. Real data is used for the design and validation of the models. The model obtained has the ability to predict water levels along the channel with minimum error, which can result in significant reduction of wasted water when implementing an automatic controller.

Efficient Operations at Intermodal Terminals Using a Multi-agent System

Transport networks are large-scale complex systems whose objective is to deliver cargo at a specific time and at a specific location. Ports and intermodal container terminals behave as exchange hubs where containers are moved from a transport modality to a different one. Terminal operations management arise as a need to face the exponentially growth of the container traffic in the last few years. In this paper the Extended Formulation of the MPC is presented. This formulation accounts for the variation of the control action to reduce not only the amount of actions but to perform a wise and efficient use of handling resources. This formulation is based on the decomposition of the control action. The Extended Formulation is applied to a simulation case study based on a long-term scheduled scenario and compared with the Basic Formulation.