Network optimization models

An application case study of the production network optimization model is reported in Chapter 5. In this context the integration of the optimization model into a planning tool to support interactive explorations of the solution space is demonstrated and guidance on how to develop the data required for quantitative strategic network design analyses is provided. Additionally, important analyses that can be performed using the proposed optimization model are introduced and improvement potentials identified in the course of a pilot application in industry are explained. 

Table 4. Classification of supply network optimization models...

To put every scenario on a common basis, CGR Management Consultants and SFI developed a common cost model. Its purpose was to eliminate these differences when evaluating the underlying changes in location. The model, called SITELINK, used a structure that is adaptable for supply chain network optimization modeling. Figure 44.1 is a simplified stmcture of the SITELINK model. 

Network optimization models are useful for managers considering regional configmation during Phase II. The first step is to collect the data in a form that can be used for a quantitative model. For SunOil, the vice president of supply chain decides to view the worldwide demand in terms of five regions— North America, South America, Europe, Africa, and Asia. The data collected are shown in Figure 5-3. 

THE CAPACITATED PLANT LOCATION MODEL The capacitated plant location network optimization model requires the following inputs ... 

We illustrate the relevant network optimization models using the example of TelecomOne and HighOptic, two manufacturers of telecommunication equipment. TelecomOne has focused on the eastern half of the United States. It has manufacturing plants located in Baltimore, Memphis, and Wichita and serves markets in Atlanta, Boston, and Chicago. HighOptic has targeted the western half of the United States and serves markets in Denver, Omaha, and Portland from plants located in Cheyenne and Salt Lake City. 

Use optimization for facility location and capacity allocation decisions. Gravity location models identify a location that minimizes inbound and outbound transportation costs. They are simple to implement but do not account for other important costs. Network optimization models can include contribution margins, taxes, tariffs, production, transportation, and inventory costs and are used to maximize profitability. These models are useful when locating facilities, allocating capacity to facilities, and allocating markets to facilities. 

Figure 28.2 Diagram of a genetic algorithm linked to a neural network for modeling and optimization.

Development and Application of Network Structure Models to Optimization of Bake Conditions for Thermoset Coatings... 

In order to answer these questions, the kinetic and network structure models were used in conjunction with a nonlinear least squares optimization program (SIMPLEX) to determine cure response in "optimized ovens ". Ovens were optimized in two different ways. In the first the bake time was fixed and oven air temperatures were adjusted so that the crosslink densities were as close as possible to the optimum value. In the second, oven air temperatures were varied to minimize the bake time subject to the constraint that all parts of the car be acceptably cured. Air temperatures were optimized for each of the different paints as a function of different sets of minimum and maximum heating rate constants. 

A network structure model has been developed from which a parameter that correlates well with physical measures of paint cure can be calculated. This model together with a kinetic model of crosslinking as a function of time and temperature has been used to evaluate the cure response of enamels in automotive assembly bake ovens. It is found that cure quality (as measured by the number and severity of under and overbakes) is good for a conventional low solids enamel. These results are in agreement with physical test results. Use of paints with narrower cure windows is predicted to result in numerous, severe under and over bakes. Optimization studies using SIMPLEX revealed that narrow cure window paints can be acceptably cured only if the bake time is increased or if the minimum heating rate on the car body is increased. 

It is the objective of this paper to provide a comprehensive review of the state-of-the art of short-term batch scheduling. Our aim is to provide answers to the questions posed in the above paragraph. The paper is organized as follows. We first present a classification for scheduling problems of batch processes, as well as of the features that characterize the optimization models for scheduling. We then discuss representative MILP optimization approaches for general network and sequential batch plants, focusing on discrete and continuous-time models. Computational... 

Several simulation-based optimization models in the context of supply chain management can be found e.g. in the area of supply chain network optimization (Preusser et al. 2005) or to simulate rescheduling of production facing demand uncertainty or unplanned shut-downs (Tang/Grubbstrom 2002 Neuhaus/Giinther 2006). A basic approach of simulation-based optimization is presented by Preusser et al. 2005, p. 98 illustrated in fig. 24. 

Three model areas can be distinguished. Model areas allow defining separated areas in the value chain network to be optimized separately. Model areas can be defined by clustered resources and/or products that have clear interfaces. Defining model area eases the implementation of a comprehensive value chain planning optimization model for a complex value chain network the optimization model can be tested for parts of the value chain network with limited data complexity before extending the model to the entire network. Three model areas are defined in the industry case study ... 

The major advantage of the prepared neural network FCC model is that it does not require a lot of input information. In addition, the model can be updated whenever new input-output information for the FCC unit is made available. This can be done by retraining the neural network starting from the old connection weights as an initial guess for the optimization process and by including the new set of data within the overall set used to train the network. 

In this chapter we presented an MILP deterministic planning model for the optimization of a petrochemical network. The optimization model presents a tool... 

Remark 1 Note that the borderlines between the three main approaches are not necessarily distinct. For instance, the targets in (ii) can be viewed as heuristics or rales that simplify the combinatorial problem and allow for its decomposition into smaller, more tractable problems (see chapter on heat exchanger network synthesis via decomposition approaches). The optimization approach (iii) can formulate thermodynamic targets, or targets on the attainable region of reaction mechanisms as optimization models, and can either utilize them so as to decompose the large-scale problem or follow a simultaneous approach that treats the full-scale mathematical model. The first... 

An optimal solution of such an optimization model will provide information on all unknown flow rates and temperatures of the streams in the superstructure (i.e., it will automatically determine the network configuration), as well as the optimal areas of the exchangers for a minimum investment cost network. 

In this section we will focus on addressing the difficulties arising in the first two tasks of sequential HEN synthesis and we will discuss simultaneous approaches developed in the early 90s. More specifically, in section 8.5.1 we will discuss the simultaneous consideration of minimum number of matches and minimum investment cost network derivation. In section 8.5.2 we will discuss the pseudo-pinch concept and its associated simultaneous synthesis approach. In section 8.5.3, we will present an approach that involves no decomposition and treats HRAT as an explicit optimization variable. Finally, in section 8.5.4, we will discuss the development of alternative simultaneous optimization models for heat integration which address the same single-task HEN problem as the approach of section 8.5.3. 

---