3 <title>Semi-Internal Processes</title>
4 <link rel="stylesheet" type="text/css" href="pythia.css"/>
5 <link rel="shortcut icon" href="pythia32.gif"/>
9 <h2>Semi-Internal Processes</h2>
11 Normally users are expected to implement new processes via the
12 <a href="LesHouchesAccord.html" target="page">Les Houches Accord</a>. Then
13 you do all flavour, colour and phase-space selection externally,
14 before your process-level events are input for further processing
15 by PYTHIA. However, it is also possible to implement a
16 new process in exactly the same way as the internal PYTHIA
17 ones, thus making use of the internal phase space selection machinery
18 to sample an externally provided cross-section expression.
19 The MadGraph 5 program [<a href="Bibliography.html" target="page">Alw11</a>] allows you to do exactly that,
20 i.e. it can be used to generate C++ code that can be linked into
21 the existing PYTHIA framework, see
22 <a href="MadGraph5Processes.html" target="page">here</a>.
25 Should you decide to go ahead on your own,
26 this page gives a brief summary how to do that. If you additionally
27 want to introduce a new resonance species, with its own internal
28 width calculations, you will find further instructions
29 <a href="SemiInternalResonances.html" target="page">here</a>. It is strongly
30 recommended to shop around for a similar process that has already
31 been implemented, and to use that existing code as a template.
32 Look for processes with the same combinations of incoming flavours
33 and colour flows, rather than the shape of the cross section itself.
34 With a reasonable such match the task should be of medium difficulty,
35 without it more demanding.
38 PYTHIA is rather good at handling the phase space of
39 <i>2 -> 1</i> and <i>2 -> 2</i> processes, is more primitive for
40 <i>2 -> 3</i> ones and does not at all address higher multiplicities.
41 This limits the set of processes that you can implement in this
42 framework. The produced particles may be resonances, however, so it is
43 possible to end up with bigger "final" multiplicities through sequential
44 decays, and to include further matrix-element weighting in those decays.
47 There are three steps involved in implementing a process:
49 <li>making use of the PYTHIA-provided kinematics information to
50 calculate the relevant cross section,</li>
51 <li>writing a new class, where the matrix elements are implemented,
52 including information on incoming and outgoing flavours and colours,
54 <li>making the process available.</li>
56 We consider these aspects in turn. An example where it all comes
57 together is found in <code>main22.cc</code>.
59 <h3>The Cross Section Calculation</h3>
61 The key method for the cross section calculation is
62 <code>SigmaProcess::sigmaHat()</code>, described below. At the point when
63 it is called, the kinematics has already been set up, and from these
64 phase space variables the differential cross section is to be calculated.
67 For a <i>2 -> 1</i> process, the returned value should be
68 <i>sigmaHat(sHat)</i>, where <code>mH</code> (= <i>mHat</i>),
69 <code>sH</code> (= <i>sHat</i>) and <code>sH2</code> (= <i>sHat^2</i>)
70 are available to be used. Incoming partons are massless. Overload the
71 <code>convertM2()</code> method below if you instead plan to return
75 For a <i>2 -> 2</i> process, instead <i>d(sigmaHat)/d(tHat)</i>
76 should be returned, based on provided
77 <code>mH, sH, sH2, tH, tH2, uH, uH2, m3, s3, m4, s4</code> and
78 <code>pT2</code> values (<code>s3 = m3*m3</code> etc.). Incoming
79 partons are massless. Overload the <code>convertM2()</code> method
80 below if you instead plan to return <i>|M|^2</i>.
83 For a <i>2 -> 3</i> process, instead <i>|M|^2</i> should be
84 returned, with normalization such that <i>|M|^2 / (2 sHat)</i> integrated
85 over the three-body phase space gives the cross section. Here no standard
86 set of Mandelstam-style variables exists. Instead the obvious ones,
87 <code>mH, sH, m3, s3, m4, s4, m5, s5</code>, are complemented by the
88 four-vectors <code>p3cm, p4cm, p5cm</code>, from which further invariants
89 may be calculated. The four-vectors are defined in the CM frame of the
90 subcollision, with massless incoming partons along the <i>+-z</i> axis.
93 In either case, <i>alpha_s</i> and <i>alpha_em</i> have already
94 been calculated, and are stored in <code>alpS</code> and <code>alpEM</code>.
95 Also other standard variables may be used, like
96 <code>CoupEW::sin2thetaW()</code>, and related flavour-dependent
97 vector and axial couplings in <code>CoupEW</code> and CKM combinations
101 In case some of the final-state particles are resonances, their
102 squared masses have already been selected according to a Breit-Wigner
103 with a linearly running width <i>Gamma(m) = Gamma(m_0) * m / m_0</i>.
104 More precisely, the mass spectrum is weighted according to
105 <i>w_BW(m^2) d(m^2)</i>, where
107 w_BW(m^2) = (1/pi) * (m * Gamma(m)) / ( (m^2 - m_0^2)^2 + (m * Gamma(m))^2 ) .
109 If you would like to have another expression, the above weights are stored
110 in <code>runBW3</code>, <code>runBW4</code> and <code>runBW5</code>,
111 respectively. If you divide out one of these factors, you just remain with
112 a phase space selection <i>d(m^2)</i> for this particle,
113 and can multiply on your desired shape factor instead. Unfortunately, the
114 Monte Carlo efficiency will drop if your new mass distribution differs
115 dramatically from the input one. Therefore it does make sense to adjust the
116 database value of the width to be slightly (but not too much) broader
117 than the distribution you have in mind. Also note that, already by default,
118 the wings of the Breit-Wigner are oversampled (with a compensating lower
119 internal weight) by partly sampling like <i>(a + b/m^2 + c/m^4) d(m^2)</i>,
120 where the last term is only used for <i>gamma^*/Z^0</i>.
123 As alternative to the kinematics variables defined above, also the two
124 arrays <code>mME[5]</code> and <code>pME[5]</code>, for masses and
125 four-momenta, respectively, can be used for cross-section calculations.
126 Here indices 0 and 1 are the two incoming beams, and 2 and onwards the
127 outgoing particles. Note that this differs by one step from the normal
128 internal labeling, where slot 0 is left empty. The four-momenta are
129 defined in the rest frame of the subcollision, with the incoming partons
130 along the <i>+-z</i> direction. The kinematics need not agree with the
131 "correct" one stored in the event record, for three reasons.
132 <br/>1) Gauge invariance forces matrix-element calculations to use
133 the same masses for incoming as outgoing legs of a particle species,
134 say <i>b</i> quarks. Therefore the kinematics of the two incoming
135 partons is recalculated, relative to the normal event record, to put
136 the partons on the mass shell. (Note that initial masses is a technical
137 issue, not the correct physics picture: the incoming partons are likely
138 to be spacelike virtual rather than on the mass shell.)
139 <br/>2) In principle each fermion flavour has to be treated separately,
140 owing to a different mass. However, in many cases fermions can be
141 assumed massless, which speeds up the calculations, and further gains
142 occur if then different flavours can use the same cross-section
143 expression. In MadGraph the default is that fermions up to and including
144 the <i>c</i> quark and the <i>mu</i> lepton are considered massless,
145 while the <i>b</i> quark and the <i>tau</i> lepton are considered
146 massive. This can be modified however, and below we provide four flags
147 that can be used to consider the "borderline" fermions either as
148 massless or as massive when matrix elements are evaluated, to match the
149 assumptions made for the matrix elements themselves.
150 <br/>3) For <i>2 -> 2</i> and <i>2 -> 3</i> processes of massive
151 identical particles (or antiparticles) in the final state, such as
152 <i>t tbar</i> or <i>W^+ W^-</i>, the kinematics is here adjusted
153 so that the two or three particles have the same mass, formed as a
154 suitable average of the actual Breit-Wigner-distributed masses. This
155 allows the evaluation of matrix-element expressions that only have
156 meaning if the two/three have the same mass.
157 <br/>Thus the mass array <code>mME[5]</code> and the four-momentum array
158 <code>pME[5]</code> present values both for initial- and final-state
159 particles based on these mass principles suited for matrix-element input.
160 Note that these variables therefore differ from the kinematics stored in
161 the event record proper, where incoming fermions are always massless and
162 outgoing resonances have independent Breit-Wigner mass distributions.
163 <br/>The conversion from the normal to the special kinematics is done
164 by calling the <code>setupForME()</code> method. This you have to do
165 yourself in the <code>SigmaHat()</code> member of your derived class.
166 Alternatively it could be done in <code>SigmaKin()</code>, i.e. before
167 the loop over incoming flavours, but then these would be considered
168 massless. The identity of final-state particles is obtained from the
169 <code>id3Mass()</code>, <code>id4Mass()</code> and <code>id5Mass()</code>
170 methods. Should the conversion to <code>mME[5]</code> and
171 <code>pME[5]</code> not work, <code>setupForME()</code> will return
172 <code>false</code>, and then the cross section should be put zero.
174 <p/><code>flag </code><strong> SigmaProcess:cMassiveME </strong>
175 (<code>default = <strong>off</strong></code>)<br/>
176 Let the <i>c</i> quark be massive or not in the kinematics set up for
177 external matrix-element evaluation.
180 <p/><code>flag </code><strong> SigmaProcess:bMassiveME </strong>
181 (<code>default = <strong>on</strong></code>)<br/>
182 Let the <i>b</i> quark be massive or not in the kinematics set up for
183 external matrix-element evaluation.
186 <p/><code>flag </code><strong> SigmaProcess:muMassiveME </strong>
187 (<code>default = <strong>off</strong></code>)<br/>
188 Let the <i>mu</i> lepton be massive or not in the kinematics set up for
189 external matrix-element evaluation.
192 <p/><code>flag </code><strong> SigmaProcess:tauMassiveME </strong>
193 (<code>default = <strong>on</strong></code>)<br/>
194 Let the <i>tau</i> lepton be massive or not in the kinematics set up for
195 external matrix-element evaluation.
199 <h3>The Cross Section Class</h3>
201 The matrix-element information has to be encoded in a new class.
202 The relevant code could either be put before the main program in the
203 same file, or be stored separately, e.g. in a matched pair
204 of <code>.h</code> and <code>.cc</code> files. The latter may be more
205 convenient, in particular if the cross sections are lengthy, or if you
206 intend to build up your own little process library, but of course
207 requires that these additional files are correctly compiled and linked.
210 The class has to be derived either from
211 <code>Sigma1Process</code>, for <i>2 -> 1</i> processes, from
212 <code>Sigma2Process</code>, for <i>2 -> 2</i> ones, or from
213 <code>Sigma3Process</code>, for <i>2 -> 3</i> ones. (The
214 <code>Sigma0Process</code> class is used for elastic, diffractive
215 and minimum-bias events, and is not recommended for use beyond that.)
216 These are in their turn derived from the <code>SigmaProcess</code>
220 The class can implement a number of methods. Some of these are
221 compulsory, others strongly recommended, and the rest are to be
222 used only when the need arises to override the default behaviour.
226 A <b>constructor</b> for the derived class obviously must be available.
227 Here you are quite free to allow a list of arguments, to set
228 the parameters of your model, or even to create a set of closely
229 related but distinct processes. For instance, <i>g g -> Q Qbar</i>,
230 <i>Q = c</i> or <i>b</i>, is only coded once, and then the
231 constructor takes the quark code (4 or 5) as argument,
232 to allow the proper amount of differentiation.
235 A <b>destructor</b> is only needed if you plan to delete the process
236 before the natural end of the run, and require some special behaviour
237 at that point. If you call such a destructor you will leave a pointer
238 dangling inside the <code>Pythia</code> object you gave it in to,
239 if that still exists.
241 <a name="method1"></a>
242 <p/><strong>void SigmaProcess::initProc() </strong> <br/>
243 is called once during initialization, and can then be used to set up
244 parameters, such as masses and couplings, and perform calculations
245 that need not be repeated for each new event, thereby saving time.
246 This method needs not be implemented, since in principle all
247 calculations can be done in <code>sigmaHat</code> below.
250 <a name="method2"></a>
251 <p/><strong>void SigmaProcess::sigmaKin() </strong> <br/>
252 is called once a kinematical configuration has been determined, but
253 before the two incoming flavours are known. This routine can therefore
254 be used to perform calculations that otherwise might have to be repeated
255 over and over again in <code>sigmaHat</code> below. For instance
256 a flavour-independent cross section calculation for a <i>q g</i>
257 initial state would be repeated 20 times in <code>sigmaHat</code>,
258 five times for the five quark flavours allowed in the incoming beams,
259 times twice to include antiquarks, times twice since the (anti)quark
260 could be in either of the two beams. You could therefore calculate the
261 result once only and store it as a private data member of the class.
262 It is optional whether you want to use this method, however, or put
263 everything in <code>sigmaHat</code>.
266 <a name="method3"></a>
267 <p/><strong>double SigmaProcess::sigmaHat() </strong> <br/>
268 is the key method for cross section calculations and returns a cross section
269 value, as described in the previous section. It is called when also a
270 preliminary set of incoming flavours has been picked, in addition to the
271 kinematical ones already available for <code>sigmaKin</code>.
272 Typically <code>sigmaHat</code> is called inside a loop over all allowed
273 incoming flavour combinations, stored in <code>id1</code> and
274 <code>id2</code>, with fixed kinematics, as already illustrated above.
275 The sum over the different flavour combinations provides the total
276 cross section, while their relative size is used to make a selection of
277 a specific incoming state.
280 <a name="method4"></a>
281 <p/><strong>bool SigmaProcess::setupForME() </strong> <br/>
282 to be called by the user from inside <code>sigmaHat()</code>
283 (or possibly <code>sigmaKin()</code>) to setup alternative kinematics
284 in the <code>mME[5]</code> and <code>pME[5]</code> arrays, better
285 suited for matrix-element calculations. See the end of the previous
286 section for a more detailed description. Should the method return
287 <code>false</code> then the conversion did not work, and
288 <code>sigmaHat()</code> (or <code>sigmaKin()</code>) should be set to
292 <a name="method5"></a>
293 <p/><strong>void SigmaProcess::setIdColAcol() </strong> <br/>
294 is called only once an initial state and a kinematical configuration has
295 been picked. This routine must set the complete flavour information and
296 the colour flow of the process. This may involve further random choices,
297 between different possible final-state flavours or between possible
298 competing colour flows. Private data members of the class may be used to
299 retain some information from the previous steps above.
300 <br/>When this routine is called the two incoming flavours have already
301 been selected and are available in <code>id1</code> and <code>id2</code>,
302 whereas the one, two or three outgoing ones either are fixed for a given
303 process or can be determined from the instate (e.g. whether a <i>W^+</i>
304 or <i>W^-</i> was produced). There is also a standard method in
305 <code>VCKM</code> to pick a final flavour from an initial one with CKM
306 mixing. Once you have figured out the value of
307 <code>id3</code> and, the case being, <code>id4</code> and
308 <code>id5</code>, you store these values permanently by a call
309 <code>setId( id1, id2, id3, id4, id5)</code>, where the last two may be
310 omitted if irrelevant.
311 <br/>Correspondingly, the colours are stored with
312 <code>setColAcol( col1, acol1, col2, acol2, col3, acol3, col4, acol4,
313 col5, acol5)</code>, where the final ones may be omitted if irrelevant.
314 Les Houches style colour tags are used, but starting with number 1
315 (and later shifted by the currently requested offset). The
316 input is grouped particle by particle, with the colour index before the
317 anticolour one. You may need to select colour flow dynamically, depending
318 on the kinematics, when several distinct possibilities exist. Trivial
319 operations, like swapping colours and anticolours, can be done with
321 <br/>When the <code>id3Mass()</code> and <code>id4Mass()</code>
322 methods have been used, the order of the outgoing particles may be
323 inconsistent with the way the <i>tHat</i> and <i>uHat</i>
324 variables have been defined. A typical example would be a process like
325 <i>q g -> q' W</i> with <i>tHat</i> defined between incoming and
326 outgoing quark, but where <code>id3Mass() = 24</code> and so the
327 process is to be stored as <i>q g -> W q'</i>. One should then put
328 the variable <code>swapTU = true</code> in <code>setIdColAcol()</code>
329 for each event where the <i>tHat</i> and <i>uHat</i> variables
330 should be swapped before the event kinematics is reconstructed. This
331 variable is automatically restored to <code>false</code> for each new
335 <a name="method6"></a>
336 <p/><strong>double SigmaProcess::weightDecayFlav( Event& process) </strong> <br/>
337 is called to allow a reweighting of the simultaneous flavour choices of
338 resonance decay products. Is currently only used for the
339 <i>q qbar -> gamma*/Z^0 gamma*/Z^0</i> process, and will likely not
340 be of interest for you.
343 <a name="method7"></a>
344 <p/><strong>double SigmaProcess::weightDecay( Event& process, int iResBeg, int iResEnd) </strong> <br/>
345 is called when the basic process has one or several resonances, after each
346 set of related resonances in <code>process[i]</code>,
347 <code>iResBeg</code> <= <code>i </code> <= <code>iResEnd</code>,
348 has been allowed to decay. The calculated weight, to be normalized
349 to the range between 0 and 1, is used to decide whether to accept the
350 decay(s) or try for a new decay configuration. The base-class version of
351 this method returns unity, i.e. gives isotropic decays by default.
352 This method may be called repeatedly for a single event. For instance, in
353 <i>q qbar -> H^0 Z^0</i> with <i>H^0 -> W^+ W^-</i>, a first call
354 would be made after the <i>H^0</i> and <i>Z^0</i> decays, and then
355 depend only on the <i>Z^0</i> decay angles since the <i>H^0</i>
356 decays isotropically. The second call would be after the <i>W^+ W^-</i>
357 decays and then involve correlations between the four daughter fermions.
360 <a name="method8"></a>
361 <p/><strong>string SigmaProcess::name() </strong> <br/>
362 returns the name of the process, as you want it to be shown in listings.
365 <a name="method9"></a>
366 <p/><strong>int SigmaProcess::code() </strong> <br/>
367 returns an integer identifier of the process. This has no internal function,
368 but is only intended as a service for the user to rapidly (and hopefully
369 uniquely) identify which process occurred in a given event. Numbers below
370 10000 are reserved for internal PYTHIA use.
373 <a name="method10"></a>
374 <p/><strong>string SigmaProcess::inFlux() </strong> <br/>
375 this string specifies the combinations of incoming partons that are
376 allowed for the process under consideration, and thereby which incoming
377 flavours <code>id1</code> and <code>id2</code> the <code>sigmaHat()</code>
378 calls will be looped over. It is always possible to pick a wider flavour
379 selection than strictly required and then put to zero cross sections in
380 the superfluous channels, but of course this may cost some extra execution
381 time. Currently allowed options are:
382 <br/>* <code>gg</code>: two gluons.
383 <br/>* <code>qg</code>: one (anti)quark and one gluon.
384 <br/>* <code>qq</code>: any combination of two quarks, two antiquarks or
385 a quark and an antiquark.
386 <br/>* <code>qqbarSame</code>: a quark and its antiquark;
387 this is a subset of the above <code>qq</code> option.
388 <br/>* <code>ff</code>: any combination of two fermions, two antifermions
389 or a fermion and an antifermion; is the same as <code>qq</code> for
390 hadron beams but also allows processes to work with lepton beams.
391 <br/>* <code>ffbarSame</code>: a fermion and its antifermion; is the
392 same as <code>qqbarSame</code> for hadron beams but also allows processes
393 to work with lepton beams.
394 <br/>* <code>ffbarChg</code>: a fermion and an antifermion that combine
396 <br/>* <code>fgm</code>: a fermion and a photon (gamma).
397 <br/>* <code>ggm</code>: a gluon and a photon.
398 <br/>* <code>gmgm</code>: two photons.
401 <a name="method11"></a>
402 <p/><strong>bool SigmaProcess::convert2mb() </strong> <br/>
403 it is assumed that cross sections normally come in dimensions such that
404 they, when integrated over the relevant phase space, obtain the dimension
405 GeV^-2, and therefore need to be converted to mb. If the cross section
406 is already encoded as mb then <code>convert2mb()</code> should be
407 overloaded to instead return <code>false</code>.
410 <a name="method12"></a>
411 <p/><strong>bool SigmaProcess::convertM2() </strong> <br/>
412 it is assumed that <i>2 -> 1</i> cross sections are encoded as
413 <i>sigmaHat(sHat)</i>, and <i>2 -> 2</i> ones as
414 <i>d(sigmaHat)/d(tHat)</i> in the <code>SigmaProcess::sigmaHat()</code>
415 methods. If <code>convertM2()</code> is overloaded to instead return
416 <code>true</code> then the return value is instead assumed to be the
417 squared matrix element <i>|M|^2</i>, and
418 <code>SigmaProcess::sigmaHatWrap(...)</code> converts to
419 <i>sigmaHat(sHat)</i> or <i>d(sigmaHat)/d(tHat)</i>, respectively.
420 This switch has no effect on <i>2 -> 3</i> processes, where
421 <i>|M|^2</i> is the only allowed input anyway.
424 <a name="method13"></a>
425 <p/><strong>int SigmaProcess::id3Mass() </strong> <br/>
427 <strong>int SigmaProcess::id4Mass() </strong> <br/>
429 <strong>int SigmaProcess::id5Mass() </strong> <br/>
430 are the one, two or three final-state flavours, where masses are to be
431 selected before the matrix elements are evaluated. Only the absolute value
432 should be given. For massless particles, like gluons and photons, one need
433 not give anything, i.e. one defaults to 0. The same goes for normal light
434 quarks, where masses presumably are not implemented in the matrix elements.
435 Later on, these quarks can still (automatically) obtain constituent masses,
436 once a <i>u</i>, <i>d</i> or <i>s</i> flavour has been selected.
439 <a name="method14"></a>
440 <p/><strong>int SigmaProcess::resonanceA() </strong> <br/>
442 <strong>int SigmaProcess::resonanceB() </strong> <br/>
443 are the codes of up to two <i>s</i>-channel resonances contributing to
444 the matrix elements. These are used by the program to improve the phase-space
445 selection efficiency, by partly sampling according to the relevant
446 Breit-Wigner distributions. Massless resonances (the gluon and photon) need not be
450 <a name="method15"></a>
451 <p/><strong>bool SigmaProcess::isSChannel() </strong> <br/>
452 normally the choice of renormalization and factorization scales in
453 <i>2 -> 2</i> and <i>2 -> 3</i> processes is based on the assumption
454 that <i>t</i>- and <i>u</i>-channel exchanges dominates the
455 cross section. In cases such as <i>f fbar -> gamma* -> f' fbar'</i> a
456 <i>2 -> 2</i> process actually ought to be given scales as a
457 <i>2 -> 1</i> one, in the sense that it proceeds entirely through
458 an <i>s</i>-channel resonance. This can be achieved if you override the
459 default <code>false</code> to return <code>true</code>. See further the
460 page on <a href="CouplingsAndScales.html" target="page">couplings and scales</a>.
463 <a name="method16"></a>
464 <p/><strong>int SigmaProcess::idSChannel() </strong> <br/>
465 normally no intermediate state is shown in the event record for
466 <i>2 -> 2</i> and <i>2 -> 3</i> processes. However, in case
467 that <code>idSChannel</code> is overloaded to return a nonzero value,
468 an intermediate particle with that identity code is inserted into the
469 event record, to make it a <i>2 -> 1 -> 2</i> or <i>2 -> 1 -> 3</i>
470 process. Thus if both <code>isSChannel</code> and <code>idSChannel</code>
471 are overloaded, a process will behave and look like it proceeded through
472 a resonance. The one difference is that the implementation of the
473 matrix element is not based on the division into a production and a
474 decay of an intermediate resonance, but is directly describing the
475 transition from the initial to the final state.
478 <a name="method17"></a>
479 <p/><strong>int SigmaProcess::isQCD3body() </strong> <br/>
480 there are two different 3-body phase-space selection machineries,
481 of which the non-QCD one is default. If you overload this method
482 instead the QCD-inspired machinery will be used. The differences
483 between these two is related to which
484 <a href="PhaseSpaceCuts.html" target="page">phase space cuts</a>
485 can be set, and also that the QCD machinery assumes (almost) massless
489 <a name="method18"></a>
490 <p/><strong>int SigmaProcess::idTchan1() </strong> <br/>
492 <strong>int SigmaProcess::idTchan2() </strong> <br/>
493 the non-QCD <i>2 -> 3</i> phase space selection machinery is rather
494 primitive, as already mentioned. The efficiency can be improved in
495 processes that proceed though <i>t</i>-channel exchanges, such as
496 <i>q qbar' -> H^0 q qbar'</i> via <i>Z^0 Z^0</i> fusion, if the identity
497 of the <i>t</i>-channel-exchanged particles on the two side of the
498 event are provided. Only the absolute value is of interest.
501 <a name="method19"></a>
502 <p/><strong>double SigmaProcess::tChanFracPow1() </strong> <br/>
504 <strong>double SigmaProcess::tChanFracPow2() </strong> <br/>
505 in the above kind of <i>2 -> 3</i> phase-space selection, the
506 sampling of <i>pT^2</i> is done with one part flat, one part weighted
507 like <i>1 / (pT^2 + m_R^2)</i> and one part like
508 <i>1 / (pT^2 + m_R^2)^2</i>. The above values provide the relative
509 amount put in the latter two channels, respectively, with the first
510 obtaining the rest. Thus the sum of <code>tChanFracPow1()</code> and
511 <code>tChanFracPow2()</code> must be below unity. The final results
512 should be independent of these numbers, but the Monte Carlo efficiency
513 may be quite low for a bad choice. Here <i>m_R</i> is the mass of the
514 exchanged resonance specified by <code>idTchan1()</code> or
515 <code>idTchan2()</code>. Note that the order of the final-state
516 listing is important in the above <i>q qbar' -> H^0 q qbar'</i> example,
517 i.e. the <i>H^0</i> must be returned by <code>id3Mass()</code>,
518 since it is actually the <i>pT^2</i> of the latter two that are
519 selected independently, with the first <i>pT</i> then fixed
520 by transverse-momentum conservation.
523 <a name="method20"></a>
524 <p/><strong>bool SigmaProcess::useMirrorWeight() </strong> <br/>
525 in <i>2 -> 3</i> processes the phase space selection used here
526 involves a twofold ambiguity basically corresponding to a flipping of
527 the positions of last two outgoing particles. These are assumed equally
528 likely by default, <code>false</code>, but for processes proceeding entirely
529 through <i>t</i>-channel exchange the Monte Carlo efficiency can be
530 improved by making a preselection based on the relative propagator
531 weights, <code>true</code>.
534 <a name="method21"></a>
535 <p/><strong>int SigmaProcess::gmZmode() </strong> <br/>
536 allows a possibility to override the global mode
537 <code><a href="ElectroweakProcesses.html" target="page">WeakZ0:gmZmode</a></code>
538 for a specific process. The global mode normally is used to switch off
539 parts of the <i>gamma^*/Z^0</i> propagator for test purposes. The
540 above local mode is useful for processes where a <i>Z^0</i> really is
541 that and nothing more, such as <i>q qbar -> H^0 Z^0</i>. The default
542 value -1 returned by <code>gmZmode()</code> ensures that the global
543 mode is used, while 0 gives full <i>gamma^*/Z^0</i> interference,
544 1 <i>gamma^*</i> only and 2 <i>Z^0</i> only.
547 <h3>Access to a process</h3>
549 Once you have implemented a class, it is straightforward to make use of
550 it in a run. Assume you have written a new class <code>MySigma</code>,
551 which inherits from <code>Sigma1Process</code>, <code>Sigma2Process</code>
552 or <code>Sigma3Process</code>, which in their turn inherit from
553 <code>SigmaProcess</code>. You then create an instance of this class
554 and hand it in to a <code>pythia</code> object with
556 SigmaProcess* mySigma = new MySigma();
557 pythia.setSigmaPtr( mySigma);
559 If you have several processes you can repeat the procedure any number
560 of times. When <code>pythia.init(...)</code> is called these processes
561 are initialized along with any internal processes you may have switched on,
562 and treated in exactly the same manner. The <code>pythia.next()</code>
563 will therefore generate a mix of the different kinds of processes without
564 distinction. See also the <a href="ProgramFlow.html" target="page">Program Flow</a>
568 If the code should be of good quality and general usefulness, it would
569 be simple to include it as a permanently available process in the
570 standard program distribution. The final step of that integration ought to
571 be left for the PYTHIA authors, but here is a description of what is
575 A flag has to be defined, that allows the process to be switched on;
576 by default it should always be off. The name of the flag should be
577 chosen of the type <code>model:process</code>. Here the
578 <code>model</code> would be related to the general scenario considered,
579 e.g. <code>Compositeness</code>, while <code>process</code> would
580 specify instate and outstate, separated by a 2 (= to), e.g.
582 When several processes are implemented and "belong together" it is
583 also useful to define a <code>model:all</code> switch that affects
584 all the separate processes.
587 The flags should normally be stored in the <code>ProcessSelection.xml</code>
588 file or one of its daughters for a specific kind of processes. This is to
589 make them easily found by users. You could create and use your own
590 <code>.xml</code> file, so long as you then add that name to the
591 list of files in the <code>Index.xml</code> file. (If not,
592 the flags would never be created and the program would not work.)
595 In the <code>ProcessContainer.c</code> file, the
596 <code>SetupContainers::init()</code> method needs to be expanded to
597 create instances of the processes switched on. This code is fairly
598 repetitive, and should be easy to copy and modify from the code
599 already there. The basic structure is
600 <br/>(i) check whether a process is requested by the user and, if so,
601 <br/>(ii) create an instance of the matrix-element class,
602 <br/>(iii)create a container for the matrix element and its associated
603 phase-space handling, and
604 <br>(iv) add the container to the existing process list.
607 Two minor variations are possible. One is that a set of related
608 processes are lumped inside the the same initial check, i.e. are
609 switched on all together. The second is that the matrix-element
610 constructor may take arguments, as specified by you (see above).
611 If so, the same basic matrix element may be recycled for a set of
612 related processes, e.g. one for a composite <i>u</i> and one for
613 a composite <i>d</i>. Obviously these variations may be combined.
618 <!-- Copyright (C) 2013 Torbjorn Sjostrand -->