Benders' Decomposition Approach To Solve Network Design Problem

The method used to solve the network design problem is based on the Benders decomposition method, where the sub problem is a mixed integer programming problem. The main problem is to consider the current constraints and choose the optimal configuration $ Q $. Where $ Q $ is the capacity vector of the warehouse. When this setting is found, a cut is generated by searching for the minimum stable delivery fee for this setting. The problem of finding the minimum stable transportation costs themselves is solved by bending decomposition where the main problem search for the worst construction of fixed configuration and cutting generation subproblems is to use the fixed simple flow problem.

All MDO methods applied to LVD in the literature classify design problems into various fields (Fig. 1). In this decomposition, the trajectory is generally regarded as the black box of the optimizer and optimized in the same way as other fields. Therefore, this decomposition may not be optimal for a particular LVD problem. Another decomposition of the design problem of coupling between orbital optimization and design variable optimization seems a valuable way to improve the overall optimization process.

To formally explain the proposed stepwise decomposition formula, we first define the rocket design. Next, analyze the rocket state vector dynamics to identify the coupling between the various flight phases. With this analysis, we were able to use decomposition to propose four formulations to solve rocket design problems. 5.2.2 Decomposition of state vector dynamics In the case of consumable launch vehicles, the global trajectory consists of various phases of the flight phase separated by phases, causing discrete changes in the state vector (quality abandonment). Each stage has its own design variables and control variables, constraints, and dynamics. As shown in the figure below, the dynamics of the state vector of the rocket can be broken down according to various phases of the flight phase. (Figure 5.1): I am I I have I am I have 3

The focus of this paper is to develop an MDO method for preliminary launch rocket design research. Our contribution is to develop a new MDO formula called SWORD (stage intelligent decomposition of best rocket design) that exploits the lateral decomposition of the design process according to the flight phase of the rocket. The SWORD formula makes it possible to distribute the complexity of the problem between the optimization level of the system and the subsystem, converts the initial global optimization problem into a series of basic subproblems that can be solved more easily I will.