Programs in Physics & Physical Chemistry
|[Licence| Download | New Version Template] adaj_v1_0.gz(38 Kbytes)|
|Manuscript Title: EXCALIBUR, a Monte Carlo program to evaluate all four fermion processes at LEP 200 and beyond.|
|Authors: F.A. Berends, R. Pittau, R. Kleiss|
|Program title: EXCALIBUR|
|Catalogue identifier: ADAJ_v1_0|
Distribution format: gz
|Journal reference: Comput. Phys. Commun. 85(1995)437|
|Programming language: Fortran.|
|RAM: 170K words|
|Word size: 32|
|Keywords: Decaying vector-boson Production, All four fermion Processes, Electroweak and QCD Background, Initial state QED radiation, Multichannel Monte Carlo approach.|
Nature of problem:
Heavy vector boson production will be investigated at e+e- colliders in a wide range of energies. At LEP II, the relevant process is
e+e- -> W+W- (1) At higher energies other processes like e+e- -> Z Z, (2) e+e- -> W e nu_e, (3) e+e- -> Z e+e-, (4) e+e- -> Z nu_e bar(nu)_e, (5)become important. The detected experimental signal for all above processes is a four fermion final state. Therefore, a Monte Carlo program being able to take into account both signal and background electroweak diagrams for all four fermion processes is required. QED initial state radiation and QCD background also play an important role and have to be included.
An event generator is the most suitable choice for a program to be able to deal with the above physical problem, since each generated event is a complete description of the momenta of the produced particles and any experimental cut can be easily implemented. There are two basic difficulties. First of all the number of Feynman diagrams can be very large. Secondly, taking into account also the background diagrams, the peaking structure of the matrix element squared is very rich, so that a straightforward integration over the allowed phase space is impractical. The former problem can be solved by using spinorial techniques to compute the amplitudes and taking massless fermions. The latter requires the use of a multichannel approach, where the integration variables are generated according to distributions that approximately reproduce the peaking behaviour of the integrand, so reducing the estimated Monte Carlo error. Since one wishes to take into account all possible final state (that means to have from 3 to 144 different Feynman diagrams, many of them leading to different peaks in the phase space), a systematic and automatic procedure for both the generation of the Feynman diagrams and the phase space integration is unavoidable, together with an algorithm for the self-optimization of the predetermined probabilities used to choose the various channels. All that has been implemented in EXCALIBUR. This paper serves also as an example of the entire procedure to be used to build future event generators.
about 100 events per second on HP, depending on the chosen physical process.
|Disclaimer | ScienceDirect | CPC Journal | CPC | QUB|