Cookies Policy
The website need some cookies and similar means to function. If you permit us, we will use those means to collect data on your visits for aggregated statistics to improve our service. Find out More
Accept Reject
  • Menu
Interest
Topics
Details

Details

  • Name

    João Pedro Pedroso
  • Role

    External Research Collaborator
  • Since

    02nd January 2006
003
Publications

2023

Stochastic crowd shipping last-mile delivery with correlated marginals and probabilistic constraints

Authors
Silva, M; Pedroso, JP; Viana, A;

Publication
EUROPEAN JOURNAL OF OPERATIONAL RESEARCH

Abstract
In this work, we study last-mile delivery with the option of crowd shipping. A company uses occasional drivers to complement its fleet in the activity of delivering products to its customers. We model it as a variant of the stochastic capacitated vehicle routing problem. Our approach is data-driven, where not only customer orders but also the availability of occasional drivers are uncertain. It is assumed that marginal distributions of the uncertainty vector are known, but the joint distribution is difficult to estimate. We optimize considering a worst-case joint distribution and model with a strategic planning perspective, where we calculate an optimal a priori solution before the uncertainty is revealed. A limit on the infea-sibility of the routes due to the capacity is imposed using probabilistic constraints. We propose an extended formulation for the problem using column-dependent rows and implement a branch-price-and-cut algorithm to solve it. We also develop a heuristic approximation to cope with larger instances of the problem. Through computational experiments, we analyze the solution and performance of the implemented algorithms.

2023

A data-driven compensation scheme for last-mile delivery with crowdsourcing

Authors
Barbosa, M; Pedroso, JP; Viana, A;

Publication
COMPUTERS & OPERATIONS RESEARCH

Abstract
A recent relevant innovation in last-mile delivery is to consider the possibility of goods being delivered by couriers appointed through crowdsourcing. In this paper we focus on the setting of in-store customers delivering goods, ordered by online customers, on their way home. We assume that not all the proposed delivery tasks will necessarily be accepted, and use logistic regression to model the crowd agents' willingness to undertake a delivery. This model is then used to build a novel compensation scheme that determines reward values, based on the current plan for the professional fleet's routes and on the couriers' probabilities of acceptance, by employing a direct search algorithm that seeks to minimise the expected cost.

2023

Deep reinforcement learning for stochastic last-mile delivery with crowdshipping

Authors
Silva, M; Pedroso, JP; Viana, A;

Publication
EURO JOURNAL ON TRANSPORTATION AND LOGISTICS

Abstract
We study a setting in which a company not only has a fleet of capacitated vehicles and drivers available to make deliveries but may also use the services of occasional drivers (ODs) willing to make deliveries using their own vehicles in return for a small fee. Under such a business model, a.k.a crowdshipping, the company seeks to make all the deliveries at the minimum total cost, i.e., the cost associated with their vehicles plus the compensation paid to the ODs.We consider a stochastic and dynamic last-mile delivery environment in which customer delivery orders, as well as ODs available for deliveries, arrive randomly throughout the day, within fixed time windows.We present a novel deep reinforcement learning (DRL) approach to the problem that can deal with large problem instances. We formulate the action selection problem as a mixed-integer optimization program.The DRL approach is compared against other optimization under uncertainty approaches, namely, sample -average approximation (SAA) and distributionally robust optimization (DRO). The results show the effective-ness of the DRL approach by examining out-of-sample performance.

2023

Novel integer programming models for the stable kidney exchange problem

Authors
Klimentova, X; Biro, P; Viana, A; Costa, V; Pedroso, JP;

Publication
EUROPEAN JOURNAL OF OPERATIONAL RESEARCH

Abstract
Kidney exchange programs (KEPs) represent an additional possibility of transplant for patients suffering from end-stage kidney disease. If a patient has a willing living donor with whom the patient is not compatible, the pair recipient-donor can join a pool of incompatible pairs and, if compatibility between recipient and donor in two or more pairs exists, organs can be exchanged between them. The problem can be modelled as an integer program that in general aims at finding the pairs that should be selected for transplant such that maximum number of transplants is performed. In this paper, we consider that for each recipient there may exist a preference order over the organs that he/she can receive, since a recipient may be compatible with several donors but the level of compatibility with the recipient might vary for different donors. Under this setting, the aim is to find the maximum cardinality stable exchange, a solution where no blocking cycle exists, i.e., there is no cycle such that all recipients prefer the donor in that cycle rather than that in the exchange. For this purpose we propose four novel integer programming models based on the well-known edge and cycle formulations, and also on the position-indexed formulation. These formulations are adjusted for both finding stable and strongly stable exchanges under strict preferences and for the case when ties in preferences may exist. Further-more, we study a situation when the stability requirement can be relaxed by addressing the trade-off between maximum cardinality versus number of blocking cycles allowed in a solution. The effectiveness of the proposed models is assessed through extensive computational experiments on a wide set of in-stances. Results show that the cycle-edge and position-indexed formulations outperform the other two formulations. Another important practical outcome is that targeting strongly stable solutions has a much higher negative impact on the number of transplants (with an average reduction of up to 20% for the bigger instances), when compared to stable solutions.

2022

The Sea Exploration Problem Revisited

Authors
Dionisio, J; dos Santos, D; Pedroso, JP;

Publication
MACHINE LEARNING, OPTIMIZATION, AND DATA SCIENCE (LOD 2021), PT I

Abstract
Sea exploration is important for countries with large areas in the ocean under their control, since in the future it may be possible to exploit some of the resources in the seafloor. The sea exploration problem was presented by Pedroso et al. [13] (unpublished); we maintain most of the paper's structure, to provide the needed theoretical background and context. In the sea exploration problem, the aim is to schedule the expedition of a ship for collecting information about the resources on the seafloor. The goal is to collect data by probing on a set of carefully chosen locations, so that the information available is optimally enriched. This problem has similarities with the orienteering problem, where the aim is to plan a time-limited trip for visiting a set of vertices, collecting a prize at each of them, in such a way that the total value collected is maximum. In our problem, the score at each vertex is associated with an estimation of the level of the resource on the given surface, which is done by regression using Gaussian processes. Hence, there is a correlation among scores on the selected vertices; this is the first difference with respect to the standard orienteering problem. The second difference is the location of each vertex, which in our problem is a freely chosen point on a given surface. Results on a benchmark test set are presented and analyzed, confirming the merit of the approach proposed. In this paper, additional methods are presented, along with a small topological result and subsequent proof of the convergence of these same methods to the optimal solution, when we have instant access to the ground truth and the underlying function is piecewise continuous.

Supervised
thesis

2022

Sistemas de Deteção de Intrusão utilizando Machine Learning

Author
Mara Quintas Almeida

Institution
UP-FCUP

2022

Customer Satisfaction Surveys: a machine learning approach

Author
Margarida Coutinho Pereira Monteiro

Institution
UP-FEP

2022

Estudo Da Implementação De Um Sistema De Monitorização E Análise De Dados Na Indústria Metalomecânica

Author
JOÃO MIGUEL SOARES CARNEIRO

Institution
IPP-ISEP

2022

Optical Extreme Learning Machines: a new trend in optical computing

Author
Duarte José Fernandes da Silva

Institution
UP-FEUP

2021

Cultural Gestures and Virtual and Augmented Reality to enable Non-Experts to Explore Systemic Concepts and Complexity

Author
Luís Miguel Alves Fernandes

Institution
UP-FCUP