Final Activity Report Summary - MCNET (Monte Carlo Event Generators for High Energy Particle Physics)
Particle physics experiments such as those at the LHC are not able to directly measure the particles that are produced by the high-energy interactions of most interest. Instead, they measure the particles produced by the particles produced by the particles produced by ... by these particles. To understand this production chain they rely heavily on simulation tools, called Monte Carlo event generators. Although there are other specialised programs that simulate one or another of these steps, there are only three programs that simulate the entire chain, called general purpose event generators, called Herwig, Pythia and Sherpa, used by all high-energy collider experiments in the world.
The MCnet Research Training Network includes the authors of all three general purpose event generators. It has added tremendous value to their individual efforts, by:
- enabling a deeper understanding of each other's models' strengths and weaknesses and thereby promoting theoretical progress on improved models;
- promoting a mobility of young researchers between the projects, furthering cross-fertilisation;
- reducing duplication in straightforward, but sometimes tedious, utility code, allowing developers to concentrate more of their effort on the core physics code;
- allowing interoperability, for a much more thorough validation of the programs against each other, against experimental data and against theoretical expectations.
These have resulted in major theoretical advances in our treatment of the evolution of such events. Some components of our simulations are firmly grounded on the underlying theories of particle physics, while others are models constrained by theory but with adjustable parameters that have to be tuned to existing data in order to make predictions for future data. Another major strand of our research activity has been in developing a generator-independent and experiment-independent framework for making generator-experiment comparisons for validation and tuning purposes, through the Cedar project. This is having a significant impact on the LHC experiments, both in terms of routinely providing event generators with both best-fit and modified-assumption parameter sets, and in terms of promoting a culture in which experimental analyses are immediately implemented into this framework for direct comparison with other generators or tunes.
The networked approach has also extended to training aspects. We have trained two graduate students through split-site, dual-project PhDs. We have also run a short-term studentship programme that allows PhD students from all over the world to join one of our research groups for three to six months. These have typically been taken by experimental students using our generators, resulting in a cascading of expert knowledge out into the user community, but have also been taken by theorists developing new models to replace one or more of our components, and a small number by our own PhD students as a mechanism for internal mobility. We have provided 35 such studentships. We have run an annual school each year, and contributed to several other training events, both for the benefit of students and postdocs in our user community and our own young researchers. These are particularly popular for the hands-on practical sessions and small-group discussions with leading experts in the field. This training has been realised with the crucial participation of four Experienced Researchers (postdocs).
All appointed researchers have been involved in the planning and execution of the training programme and have been represented in the management of the network by a Student and Postdoc Committee. The final product of the network has been for our event generators to have been fully ready for the start of LHC data-taking. Network members have played a significant role in collaboration with the LHC experiments in tuning our models to the first data, evaluating the remaining modelling uncertainties, using them to quantify detector calibration issues, and ultimately contributing to the search for new physics and measurements of known physics processes.
The MCnet Research Training Network includes the authors of all three general purpose event generators. It has added tremendous value to their individual efforts, by:
- enabling a deeper understanding of each other's models' strengths and weaknesses and thereby promoting theoretical progress on improved models;
- promoting a mobility of young researchers between the projects, furthering cross-fertilisation;
- reducing duplication in straightforward, but sometimes tedious, utility code, allowing developers to concentrate more of their effort on the core physics code;
- allowing interoperability, for a much more thorough validation of the programs against each other, against experimental data and against theoretical expectations.
These have resulted in major theoretical advances in our treatment of the evolution of such events. Some components of our simulations are firmly grounded on the underlying theories of particle physics, while others are models constrained by theory but with adjustable parameters that have to be tuned to existing data in order to make predictions for future data. Another major strand of our research activity has been in developing a generator-independent and experiment-independent framework for making generator-experiment comparisons for validation and tuning purposes, through the Cedar project. This is having a significant impact on the LHC experiments, both in terms of routinely providing event generators with both best-fit and modified-assumption parameter sets, and in terms of promoting a culture in which experimental analyses are immediately implemented into this framework for direct comparison with other generators or tunes.
The networked approach has also extended to training aspects. We have trained two graduate students through split-site, dual-project PhDs. We have also run a short-term studentship programme that allows PhD students from all over the world to join one of our research groups for three to six months. These have typically been taken by experimental students using our generators, resulting in a cascading of expert knowledge out into the user community, but have also been taken by theorists developing new models to replace one or more of our components, and a small number by our own PhD students as a mechanism for internal mobility. We have provided 35 such studentships. We have run an annual school each year, and contributed to several other training events, both for the benefit of students and postdocs in our user community and our own young researchers. These are particularly popular for the hands-on practical sessions and small-group discussions with leading experts in the field. This training has been realised with the crucial participation of four Experienced Researchers (postdocs).
All appointed researchers have been involved in the planning and execution of the training programme and have been represented in the management of the network by a Student and Postdoc Committee. The final product of the network has been for our event generators to have been fully ready for the start of LHC data-taking. Network members have played a significant role in collaboration with the LHC experiments in tuning our models to the first data, evaluating the remaining modelling uncertainties, using them to quantify detector calibration issues, and ultimately contributing to the search for new physics and measurements of known physics processes.