Decision Science in Port Operations

In the context of port nautical chain/services, we observe port authorities and port ecosystem participants adopting various optimization strategies for achieving

• Sustainability goals
• Productivity and efficiency goals
• Profitability goals

Here we look at strategies that employ mathematical optimization techniques along with simulation, machine vision, GIS, IoT. These tools are deployed in conjunction with other software systems like PCS, project management tools, databases, HRMS etc.

These tools provide decision support to port operators, build and evaluate what if scenarios and generate plans for optimized operations. They aid port stakeholders to monitor and improve KPIs like turnaround, emissions, throughput etc.

Like other classical applications in supply-chain and transportation we observe the impact of decision science in resource allocation, cost reduction and automation.

Here we will explore two major approaches that we see ports across the world adopt as far as mathematical modelling is concerned vis a vis

A. Port operations modelling

B. Mathematical Optimization

Port operations modelling

The goal of “port nautical chain services” modelling tools is to provide a quantitative tool for assessment of the port performance and for ex-ante assessment of improvement measures. These models simulate the port’s system behavior, generate system KPI’s and provide insights on factors that can lead to a deterioration or improvement of the port performance. Subsequently, the improvement measures may concern operations of individual parties at the port or improvement measures may cross organizational boundaries, such as sharing real-time information and coordination between the actors in the chain

The most prevalent port operations modelling techniques are

a. Agent based model simulation

b. Discrete event simulations

c. Hybrid model (a. +b.)

d. GIS based digital Twins

These models can provide strategic decision support through scenario-wise evaluation of beneficial improvement measures.

Examples of application of Mathematical Optimization techniques in ports

We observe the use of various mathematical programming techniques like MILP (mixed integer linear programming), Branch and cut algorithm, chaos particle swarm optimization method, genetic algorithms etc. in practice. These optimizations techniques aim to achieve improvement in operations on both landside and seaside operations.

Some implementations and the goals they achieve are listed below:

Tugboat scheduling problem (TuSP) and Tug-boat Fleet management (TBFM)

a. Implementations observed by players at Singapore Port and Port of Rotterdam, Port of Antwerp

b. Dual objectives of maximizing the utilization rate of tugboats and minimizing the total waiting time of ships

c. The tugboat scheduling obtained by these tools uses fewer tugboats and incurs less travel cost. Impact (about 6-8% i.e. $150 per trip).Below is a sample of a schedule for tugboats. Note: The numbers connected by arrows are node numbers, representing different locations at the port. The two numbers under each node number are the tugboat arrival time and service start time in minutes.

Allocation and Scheduling of Quay Cranes, Yard Cranes, and Trucks in dynamically integrated Container Terminal Operations

We encounter an implementation of a dynamical modeling of integrated (end-to-end) container terminal operations using finite state machine (FSM) framework where each state machine is represented by a discrete-event system (DES) formulation. This hybrid model incorporates the operations of quay cranes (QC), internal trucks (IT), and yard cranes (YC) along with the selection of storage positions in container yard (CY) and vessel bays.

Integrated Berth and Quay Crane Allocation

To address the problem of integrated berth and quay crane allocation (I-BCAP) in general seaport container terminals, use of a model predictive allocation (MPA) algorithm and preconditioning methods are observed.

Other prominent studies have been observed in process mining. The applications nevertheless are not restricted to the ones covered in this note. Also, successful implementation and adoption of these methods depend on the present state of IT systems and processes, particularly in individual ports.