Distributed Solution of Simulation-Based Optimization Problems on Networks of Workstation

THE PAPER DEALS WITH THE DISTRIBUTED SOLUTION OF SIMULATION-BASED NONLINEAR CONSTRAINED OPTIMIZATION PROBLEMS IN ENGINEERING ON A NETWORK OF WORKSTATIONS. THIS TYPE OF NONLINEAR OPTIMIZATION AND CONTROL PROBLEMS IN ENGINEERING CAN BE GENERALLY CHARACTERIZED AS NON CONVEZ AND NON SMOOTH. ADDITIONALLY, THE INVOLVED SIMULATION INTRODUCES CERTAIN NUMERICAL NOISE TO THE SOLUTION PROCESS, BECAUSE OF LONG -RUNNING CUMPATATIONS FOR SIMULATION AND DIFICULTIES IN FINDING A GLOBAL OPTIMISMUM. THE RESULTING REQUIREMENTS FOR THE CORRESPONDING SOFTWARE INTEGRATION OF SIMULATION AND THE DIFICULTIES IN FINDING A GLOBAL OPTIMUS. THE RESULTING REQUERIMENTS FOR THE CORRESPONDING SOFTWARE ARCHITECTURE AND SOFTWARE INTEGRATION OF SIMULATION AND DIFICULTIES IN FINDING A GLOBAL OPTIMUM. THE RESULTING REQUERIMENTS FOR THE CORRESPONDING SOFTWARE ARCHITECTURE AND SOFTWARE INTEGRATION OF SIMULATION AND OPTIMIZATION , COMPONETES ARE DISCUSED THE IMPLEMENTATION OF THIS SOFTWARE ARCHITECTURA IS PRESENTED USING THE OP TEX SYSTEM FOR OPTIMIZATION, THE FINITE -ELEMENT PACKAGE FEFLOW FOR SIMULATION AND THE WINNER RESOURCE MANAGEMENT SYSTEM FOR LOAD DISTRIBUTION IN NETWORKS OF WORKINS. FURTHERMORE, THE CONCEPTS OF A DISTRIBUTED OPTIMIZATION ALGORITHM FOR LOAD DISTRIBUTION IN NETWORS OF WORSTATIONS. FURTHERMORES THE CONCEPTS OF A DISTRIBUTED OPTIMIZATION ALGORITHM ARE DESCRIBED AND PRELIMINARY RESULTS OF ITS APPLICATION TO INDUSTRIAL APPLICATIONS FROM WATER ENGINEERING ARE USED TO SHOW THE FEASIBLILITY OF THE PROPOSED SYSTEM FOR DISTRIBUTED OPTIMIZATION.