Current Research
Antonio Celio P. Mesquita
03072248800

Qualification exam
May 25, 2022

Defense date
May xx, 2024

Advisor
Carlos Alberto Alonso Sanches

Type
Doctorate
Abstract
In the scenario of aerial pickup and delivery of goods in a distribution network, transport aviation faces risks of cargo unbalancing due to the urgency required for loading for rapid take-off and mission accomplishment, especially in times of crisis, public calamity, client contract short lead times, or any external pressure for immediate take-off. We contributed by modelling and solving a new problem of planning the loading and routing of an aircraft according to a utility score, weight and balance principles, and fuel consumption in a tour of simultaneous pickup and delivery at intermediate airfields. This difficulty encompasses four sub-problems: Air Palletization; Weight and Balance; Pickup and Delivery; and Vehicle Routing. This new hard problem, named Air Cargo Load Planning with Routing, Pickup, and Delivery Problem (ACLP+RPDP), is mathematically modelled using standardised pallets in fixed positions, obeying the centre of gravity constraints, delivering each item to its destination, and minimising fuel consumption costs. We also contributed by carrying out multiple experiments with a commercial solver and four well-known meta-heuristics on synthetic data based on real data from the Brazilian Air Force's transportation. These challenging benchmark instances are made publicly available. We also created a heuristic that quickly finds good solutions for a wide range of problem sizes, an essential contribution as it was the unique method that managed to solve all testing scenarios. (222 words)
Provisional title
Air cargo load and route planning in pickup and delivery operations

Research problem
There is no technological assistance to help the load and trip planners with the huge number of demands for transport in each hub. This enables risks of cargo unbalancing, improper delivery, excessive fuel burn, and more than 2 hours of TAT for airlift missions. It is necessary to solve an Air Paletization, a Weight and Balancing, a Pickup and Delivery, and a Routing Problem, all simultaneously.
José Nogueira M. Filho
37264559734

Qualification exam
Ago. 2022

Defense date
Apr xx, 2024

Advisor
Fernando Teixeira Mendes Abrahão

Type
Doctorate
Abstract
When planning maintenance for complex systems, clients need to assure that system is safe for the operation, that it can accomplish the planned missions in a cost-effective way. The problem is that initial maintenance planning is conservative due to lack of a maturity data and tools to support the maintenance engineers that follow MSG-3 guidelines for maintenance task analysis and maintenance plan design, strictly using their best “engineering judgment”, which may result in inefficient maintenance plans, i.e., more costly than they could be, and with lower than possible availability and dispatchability. As aviation costs are one of the highest in the industry, even a small amount of savings could represent a great amount of money in the long run. This study adds to the research on complex system maintenance by looking at how an optimization model and field or usage data could be used to make maintenance plans that are smart and reliable. An optimal maintenance plan must associate tasks to maintenance packages and preparation tasks in many cycles along the airliner's life span. We extend the study on Task Allocation Problem (TAP) by evaluating the possibility of these optimizations with supervised learning on historical data to construct a resilient maintenance plan. The results proved that the model task grouping associated with machine learning reduces total maintenance costs in the long run. The results also revealed little influence of repair probability on aeronautical components in the optimization model. (238 words)
Provisional title
Data Learning guided optimization model for tasks to maintenance package allocations

Research problem
Maintenance planning is conservative due to limitations faced by maintenance engineers, such as the absence of an efficient tool to complement the MSG-3 analysis and support the allocation of tasks during the design of the operator’s maintenance plan, a Task Allocation Problem (TAP), which may result in inefficient maintenance plans, i.e., more costly than they could be, and with lower than possible availability.