Software system for solving multi-scale optimization problems [chapter]

E. Semenkin, K. Abramovich
1996 System Modelling and Optimization  
Optimization problems with different type of variables are described. Methods of decomposition and unification are suggested for solving these problems. Software system, that realizes mentioned methods, is described. Results of testing and ways of further development are discussed. EXANWLESOFPROBLEM Optimization problems with the variables of different type (discrete, boolean, continuous, combinatorial, etc.) are called multi-scale or mixed problems. They arise in various design and planning
more » ... blems when mathematical formulations and modelling are possible. Consider two examples of problems of that type. 1) CAD of some large scale system consisting of several interconnected units (subsystems). The objective F(B, D, X) is some quality index, representing the ability of the system to perform its functions. B E Bn is the vector of Boolean variables representing the system structure. If bi = 1 then i-th subsystem is included in the system and vice verse. D E Zm is the vector of integer variables, representing the number of backup copies of subsystems. X E Rk is the vector of continuous variables representing the physical characteristics of the subsystems (mass, power consumption, etc.). 2) Planning of optimal sequence of an industrial projects' implementation. J. Doležal et al. (eds.), System Modelling and Optimization
doi:10.1007/978-0-387-34897-1_60 fatcat:zzk2hwbzcnbhfed4wvb6wutpmm