- Pedro M. Castro, Ana P. Barbosa-Póvoa, Henrique A. Matos, Augusto Q. Novais. Simple Continuous-time Formulation for Short-term Scheduling of Batch and Continuous Processes. Ind. Eng. Chem. Res. 2004, 43, 105-118.
- Pedro M. Castro and Ignacio E. Grossmann. Generalized Disjunctive Programming as a Systematic Modeling Framework to Derive Scheduling Formulations. Ind. Eng. Chem. Res. 2012, 51, 5781-5792.
- João P. Teles, Pedro M. Castro and Henrique A. Matos. Multi-parametric disaggregation technique for global optimization of polynomial programming problems. Journal of Global Optimization 2013, 55, 227-251.
- Iiro Harjunkoski, Christos Maravelias, Peter Bongers, Pedro Castro, Sebastian Engell, Ignacio Grossmann, John Hooker, Carlos Méndez, Guido Sand and John Wassick. Scope for Industrial Applications of Production Scheduling Models and Solution Methods. Computers & Chemical Engineering 2014, 62, 161-193.
- Pedro M. Castro. Optimal Scheduling of Multiproduct Pipelines in Networks with Reversible Flow. Ind. Eng. Chem. Res. 2017, 56, 9638-9656.
Pedro M. Castro
Indicadores
ResearcherIDScopus
Google Scholar
Os meus interesses científicos situam-se na interface entre a Engenharia Química, Engenharia e Gestão Industrial e Investigação Operacional, sendo as áreas designadas por Engenharia de Sistemas de Processos e Optimização Discreta. Em concreto, desenvolvo modelos e algoritmos com base em programação linear inteira mista (MILP) para: (i) otimização de operações de processos industriais, ao nível do escalonamento e planeamento de produção; (ii) otimização global de problemas não-convexos com variáveis inteiras do tipo bilinear.
(i) Especialista em modelos matemáticos com diferentes conceitos de representação temporal (ex. tempo discreto e contínuo) para escalonamento de produção. Restrições complexas são incorporadas através do recurso à representação Resource-Task Network (RTN) e/ou Generalized Disjunctive Programming (GDP). Como exemplo, importa referir a modelação de custos de energia elétrica variáveis no tempo, algo importante no conceito de redes inteligentes de energia. Estive envolvido na resolução de problemas de empresas nas áreas de aeronáutica, produtos alimentares, pasta e papel, química fina, plásticos, cimento, aço e petróleo, nomeadamente com: Acciai Speciali Terni, Dow Chemical, ExxonMobil, Metsä Tissue, SASOL, Hovione, Portucel, Celulose do Caima, FIMA, Ferro Portugal.
(ii) Especialista em relaxações MILP de problemas polinomiais do tipo NLP/MINLP com base na discretização do domínio das variáveis. Juntamente com o meu ex-aluno de doutoramento João Teles, desenvolvi o método de transformação Multiparametric Disaggregation Technique. Este tem a vantagem de adicionar um número de variáveis inteiras ao problema que cresce de forma logarítmica com o número de partições, permitindo um aumento significativo da performance computacional quando comparado com outros métodos alternativos. A sua aplicação em problemas de otimização de redes de águas industriais, centrais hidroeléctricas e misturas de combustíveis, tem-se vindo a mostrar competitiva com os algoritmos comerciais de otimização global.
Fiz parte da equipa de desenvolvimento da infraestrutura web www.minlp.org, sendo o principal contribuinte para a biblioteca de modelos.
Especialização nos softwares GAMS, gPROMS, Aspen Plus, Visual Basic.
Colaborações internacionais ativas com: Carnegie Mellon University (Prof. Ignacio Grossmann); ABB Corporate Research (Dr. Iiro Harjunkoski).
My scientific interests lie at the interface between Chemical Engineering, Industrial Engineering and Operations Research. The areas are known as Process Systems Engineering and Discrete Optimization.
I develop models and algorithms based on mixed-integer linear programming for: (i) optimization of operations in industrial processes at the level of production planning and scheduling; (ii) global optimization of non-convex bilinear problems featuring binary variables.
(i) Expert in mathematical models relying on different concepts of time representation (e.g. discrete- and continuous-time) for scheduling. Complex constraints are handled by using the Resource-Task Network (RTN) process representation and/or Generalized Disjunctive Programming (GDP). As an important example in the context of smart grids, we have the incorporation of time dependent electricity costs. I have been involved in solving problems originating in companies from the fields of aeronautics, food, pulp and paper, fine chemicals, plastics, cement, steel and petroleum: Acciai Speciali Terni, Dow Chemical, ExxonMobil, Metsä Tissue, SASOL, Hovione, Portucel, Celulose do Caima, FIMA, Ferro Portugal.
(ii) Expert in MILP relaxations for NLP/MINLP polynomial problems, which rely on the discretization of the variables domain. Together with my former PhD student João Teles, I developed the transformation method Multiparametric Disaggregation Technique (MDT). It has the advantage of adding to the original problem a number of binary variables that grows logarithmically with the number of partitions, leading to improvements in performance when compared to alternative methods. The application of MDT in problems dealing with the design of industrial water networks, hydroelectric power production, and blending of crude or refined petroleum products, has been shown competitive to commercial global optimization solvers BARON and GloMIQO.
I was part of the development team for the cyberinfrastructure www.minlp.org, being the main contributor to the library of models.
Expertise in GAMS, gPROMS, Aspen Plus and Visual Basic software.
Ongoing international collaborations with Carnegie Mellon University (Prof. Ignacio Grossmann); ABB Corporate Research (Dr. Iiro Harjunkoski).