A hierarchical architecture of modelling on integrated supply chain optimization systems of continuous process industries

Towards the integration of supply chain and investment decision making, this paper proposed a hierarchical architecture of modelling on integrated supply chain optimization system, which consist of such three levels as integrated long range planning and decision making level, sustainable supply chain planning and scheduling level, unit process simulation and optimization level. At long range planning level, a goal programming (GP) model is utilized todebottleneck the supply chain. If there is some expansion needed, investment evaluation will be executed at this level basing on the analytic hierarchy process. The MIGP approach is utilized to make an optimum decision on which action or coupled actions to be selected. At supply chain midterm planning basing on GP model and short-term scheduling basing on the MIGP or MINLP model will be executed at various time horizons from annual to daily. All models on these two levels are simplified with the parameters offered by the bottom level. At the bottom level, in order to get quick solution, the ANN models combined with GA are utilized to simulate and optimize the unit operation processes under the constraints offered by the higher levels. The rigorous models will be utilized to training the ANN models. The application of this hierarchical architecture to a refinery plant is given.