We propose a global optimization algorithm for mixedinteger nonlinear programming (MINLP) problems arising from oil refinery planning. It relies on tight mixed-integer linear programming (MILP) relaxations that discretize the bilinear terms dynamically using either piecewise McCormick (PMCR) or normalized multiparametric disaggregation (NMDT). Tight relaxations help finding a feasible solution of the original problem via a local nonlinear solver, with the novelty being the generation of multiple starting points from CPLEX's solution pool and the parallel execution. We show that optimalitybased bound tightening (OBBT) is essential for large-scale problems, even though it is computationally expensive. To reduce execution times, OBBT is implemented in parallel. The results for a refinery case study, featuring units with alternative operating modes, intermediate storage tanks, and single-and multiple-period supply and demand scenarios, show that the algorithm's performance is comparable to commercial solvers BARON and ANTIGONE.
This work uses multi-period, inventory pinch-based algorithm with continuous-time model (MPIP-C algorithm 1 ) for scheduling linear or nonlinear blending processes. MPIP-C decomposes the scheduling problem into (i) approximate scheduling and (ii) detailed scheduling. Approximate scheduling model is further decomposed into two parts: a 1 st level model which optimizes nonlinear blend models (with time periods delineated by inventory pinch points), and a 2 nd level multi-period mixed-integer linear programming model (which uses fixed blend recipes from the 1 st level solution) to determine optimal production plan and swing storage allocation, while minimizing the number of blend instances and product changeovers in the swing tanks. The 3 rd level computes schedules using a continuous-time model including constraints based on the short-term plan solution. Nonlinear constraints are used for the Reid vapor pressure in our case studies. Excellent computational performance is illustrated by comparisons with previous approach with discrete-time scheduling model.
Many real-life complex networks have in-degree and out-degree distributions that decay as apower-law. However, the few models that have been able to reproduce both of these properties,cannot reproduce the wide range of values found in real systems. Another limitation of thesemodels is that they add links from nodes which are created into the network, as well as betweennodes already present in this network. However, adding links between existing nodes is not acharacteristic available in all systems. This paper introduces a new complex network growthmodel that, without adding links between existing nodes is able to generate complex topologieswith in-degree and out-degree distributions that decay as a power-law. Moreover, in this growthmodel, the ratio at which links are created is greater than the ratio at which nodes are born, whichproduces an accelerated growth phenomenon that can be found in some real systems, like theInternet at the Autonomous System level.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.