The main objective of my research is to develop systematic procedures for process systems that are relevant to the Chemical and Biochemical Process Industries, namely to develop methods and tools in Process Systems Engineering (PSE). In particular, several problems are addressed in the areas of process synthesis, planning and scheduling that rely on linear and nonlinear models with continuous and discrete decisions. Such models can be expressed within mixed integer programming (MILP and MINLP), disjunctive programming, and multi-period optimization frameworks.

On the one hand, being an area that addresses industrial problems by linking science to engineering, PSE requires continuous interaction with various fields of engineering and applied sciences, with increasing ties to the natural sciences. On the other hand, operations research and applied mathematics provide the fundamental theory and supporting tools. Moreover, advances in computing help to realize some of the supporting tools and many of the functions currently available for parallel computing will become more accessible through advanced software.

Research topics that address important problems associated with process systems engineering are described below.

(i) Supply Chain, Planning and Scheduling
Involves the development of supply chain models for process systems. Interestingly, very few groups in chemical engineering worldwide devote attention to such problems that require the investigation of several complex problems in process operations, such as planning and scheduling of batch and continuous processes, supply of raw materials and product distribution. All these elements must be integrated over time in order to form the so-called supply chains. Major challenges include the development of methods for simulating and optimizing under uncertainty. Moreover, optimization tools that can handle mixed-integer, discrete-logic and qualitative equations, algebraic and/or dynamic to model such problems more effectively are required.

Process planning and scheduling activities include the development of optimization models and solution techniques for chemical processes. Important industrial-scale development projects are currently being carried out with financial support from chemical and petrochemical companies. Major effort is being concentrated on the development of mixed integer optimization models for scheduling and planning of continuous processes, for which no significant literature is yet available.

Another important feature in this context involves transportation operations, usually considered one of the major bottlenecks in the production chain. While delays imply loss of time and lack of resources, deliveries ahead of deadlines may cause excess of inventories. Pipelines provide an economic and continuous fluid transportation mode, especially when large amounts of these products have to be pumped for long distances. Therefore short-term scheduling problems in supply and distribution complexes must be addressed.

The integration of planning and scheduling models is a particularly important issue for large systems, a typical example being an oil refinery[1]. Although full-scale planning models have been developed for refineries, only local scheduling models (crude scheduling, fuel oil/asphalt scheduling, LPG scheduling, atmospheric distillation scheduling) can be effectively solved with the current methods.

These planning and scheduling models must be properly coordinated in order to generate consistent solutions. Integration must also include raw material and distribution logistics. In the case of batch processes, performance models are being incorporated into short term scheduling models in order to account for different process conditions. On the algorithmic side, logic based[2] and metaheuristic techniques[3], as well as disjunctive approaches, are being explored in the search for computationally efficient solutions.

(ii) Process Operations
In the second main research area, methodologies for process operations are being investigated, with emphasis on problems that are modeled with constraints that are composed by differential-algebraic systems (optimal control problems). This area demands a number of new supporting methods and tools that are not available, such as large-scale methods that solve systems at multiple scales as well as improved real-time optimization methods to handle extremely large models with millions of variables. Important applications under development involve the operation of wastewater treatment stations and utility systems. A recent application that reflects a shift in process systems engineering involves the application to mathematical biology with emphasis on problems in immunology, virology, and cell and molecular biology. Recently, we have been looking into the optimal control of the viral dynamics of HIV.

Industrial facilities that operate with several products in a fixed number of production lines generate effluents with varying characteristics and multiple substrates. In this sense, optimal strategies for the operation of wastewater treatment plants must be developed, in terms of the management of equalization tanks and in the interaction between scheduling and operation. In addition, the simultaneous synthesis and design of stations will be investigated. Methodologies for the optimal operation of anaerobic digesters are being integrated in the previous problems with the objective of reducing sludge generation. Finally the objective the to maximize water reuse, by synthesizing mass exchange networks simultaneously with the operation of treatment plants.

A research field that brings a lot of enthusiasm follows the mathematical modeling directed to the pathogenesis of HIV, whose chemotherapy treatment can be posed as an optimal control problem[4]. The first step consists of the study and development of deterministic optimization models that are able to represent at the cellular level the interaction between the HIV-1 virus and the human immune system during the infection and progression of AIDS. We then must propose exact solution techniques for such models with the objective of defining an optimal strategy for the chemotherapic treatment based on a set of hypothesis inherent to the immunology and pathology of the type-1 HIV virus. This strategy should propose guidelines to control or minimize the infection in the human body.

(iii) Process Synthesis and Design
The third line of research concerns problems that are related to process synthesis and design, that are formulated as disjunctive and mixed-integer optimization models[5]. A major challenge is to apply methodologies developed previously to process design. New applications are being expanded that include bioprocesses and biomedical devices. Other areas that are likely to receive increasing attention are the design of metabolic networks and molecular design. Strategies are under development for the optimization of production and purification processes for recombinant proteins. The aim is to develop systematic and optimized procedures for the rational design of protein separation/purification sequences to isolate and purify therapeutic and other proteins from a large number of contaminants present in a production stream.

Today there are a very large number of new protein products coming into the market. Many of them are in the final stages of clinical trials that will require production facilities. Furthermore, as has been seen in the last few years this growth is exponential (e.g. Clinical Trials Search Results from Chiron and Genentech available in www (http://www.recap.com)). The demand for these proteins can vary several fold from year to year, depending on the size of the potential market (which keeps increasing with time as costs go down) and the clinical success of the product (e.g. insulin, hepatitis B vaccine, human growth hormone, tPA for heart patients). An adequate fulfillment of this demand can only be achieved by an efficient multiproduct/multipurpose batch plant like the one that will be investigated in the present project. On the other hand, the practical methodologies used to carry out such biotechnological processes are still in their infancy, hence new processes are constantly being developed (e.g. liquid-liquid extraction of proteins by Aqueous Two-Phase Systems or reversed micelles) that need to be evaluated.

Strategies for optimized design of protein production and recovery sequences will increase the protein recovery yields, reduce the separation/purification and production costs, and shorten time for bringing new protein products into production and hence use. An international cooperation project (Argentina-Brazil-Chile) supports the groups of the University of Sao Paulo, INGAR and the University of Chile that have a history of successful previous collaboration in a project that investigated the computational, mathematical and more theoretical aspects of this problem.

Apart from the aforementioned research areas, theoretical aspects related to combinatorial optimization, logic-based optimization, disjunctive programming should be investigated. A strong interaction with groups in operations research, computer science and applied mathematics is necessary. In summary, there are plenty of intellectually challenging problems in PSE that can integrate science-based research and industry[6].