《Multi-objective Genetic Optimization for LCLS-II X-ray FEL》.pdf

JournalofMechanicsEngineeringandAutomation4(2014)632-638 lN 《旨 Multi-objectiveGeneticOptimizationforLCLS·IIX_ray FEL LanfaW angandTorO．Raubenheimer SLAC(StanfordLinearAcceleratorCenter)，StanfordUniversity，California94025，USA Received：May09，2014 ／Accepted：May31，2014 ／Published：August25，2014 Abstract：TheLCLS—II(1inaccoherentlightsourceII)willbuildontheSuccessoftheworld’smostpower~lX’raylaser：theLCLS flinaccoherentlightsource)．ItwilladdtwonewX—raylaserbeamsandroom foradditionalnew instruments，thusgreatlyincreasnig thenumberofexperimentscarried outeach year．Multiple operation modesareproposed to accommodatea variety ot er reauirements．Therearealargenumberofvariablesandobjectivesinthedesign．Foreachoperationmode，MOGA(multi—objective geneticalgorithm)isappliedtooptimizethemachineparametersinordertominimizethejitters，energyspread，collectiveeffectsand emittance． Keywords：MOGA，optimization，FEL(free—electronlaser) 1．Introduction possibility of LCLS—I1 with two beam energies； Section5discussestheminimizationofCSR induced LCLS—II(1inaccoherentlightsourceII)willprovide emittancegrowth；andSection6givesconclusions． beam withdifferencecharge．Foreachbeam ，theR56 atBC1 (bunch compressor 1、andBC2 (bunch 2．BenchmarkwithLCLSBearn compressor 2)，accelerating structure phase and It is important to have a comparision of the voltageareoptimizedusingMOGA (multi-objective simulationwith themeasurements．A seriesofdata， geneticalgorithm)program to satisfy therequired