Title: Transdisciplinary Physics: application of statistical physics to social issues and risk
1Trans-disciplinary Physics application of
statistical physics to social issues and risk
- A proposal for COST action
- Peter Richmond, Ireland
- Janusz Holyst, Poland
- and
- Sorin Solomon, Israel
2Assessment of risk is of vital concern in many
areas of public interest
- Finance and insurance
- Manufacturing and marketing
- Health
- Food safety
- Public policy making
3Finance and Insurance
- Speculative bubbles and manias
- 17th Century - tulips
- 18th Century - early emerging markets
- South sea Mississippi
- 20th Century
- Wall St 1929, 1973/4, 1987, 1998/9
- 21st Century
- TMT 2000-200?
- Need better assessment of future risk
- Long Term Capital Management asset management USA
- Equitable Life UK
- Enron
4Manufacturing and marketing
- Eg Just in time for ice cream and soft drinks
manufacturing - Consumer choice product selection
5Health
6Food in Europe much unease, much concern..
- Salmonella in eggs
- Lysteria in cheese
- Benzene in Perrier water
- BSE in cattle
- Aids in blood
- Plague in pigs
- Dioxin in chicken
- The Coca Cola mystery
- Antibiotics in meat
- Foot and mouth disease
- Bayes Inverse probabilities
7Public policy making
- Expenditure and risk
- structure of public transport networks
- Public health strategy
- Social expenditure
- Defence and warfare resource allocation
- Polling and social opinion
8Research networks
9Mega amounts of new financial data since 1985
10Some stylised facts
- Power law distributions
- Clustered volatility
- Correlations
11Agents models and forces
12Forces in people agents and groups
buy
Hold
Sell
13Atoms are inert Agents can learn to .
- have memory
- persuade or convince
- act as group
- Self organise into cooperative networks
- Be heterogeneous and differ in world knowledge
and view
14New emergent area of physics?
- UK Institute of Physics Physics in Finance 2000
- German Physical Society Socio-physics 2000
- European Physical Society has already held three
conferences on Quantitative Finance (1999, 2000,
2001) 2003 Warsaw - France, Germany, US Santa Fe, Italy, Bali (?)
15Opportunity
- Use new and powerful tools to address important
questions of public interest - Risk
- finance, food safety
- Group behaviour
- polling, referendums, consumer choice
- Networks
- Facilitate new European network
- NOE in FP6 (Santa Fe)
- Expand recognition of physics and IT
- Areas of great interest to young people
16Three working groups
- WG1 Physics risk
- What does the data tell us?
- Empirical models
- Baysian belief models
- WG2 Learning agent models
- Minority games
- Evolutionary dynamics stochastic equations
- Probabalistic approaches Fokker Planck type
equations - Memory effects non Markovian processes
- WG3 Network dynamics topology
- Structure, dynamics and evolution
17Management structure timeline
18Risk COST action Benefits
- New research expertise brought to bear on issue
of public importance - Risk quantitative finance
- Risk manufacturing marketing
- Risk food safety
- New trans-disciplinary collaborations
- Social economic sciences ?? physical and IT
sciences - Bridge between innovative new member state
programmes - Stairway to new proposals
- Eg FP6 RTN?
- Increase understanding of innovative areas of
physics - Microscopic models interacting agents learning
models networks
19Physics and risk
- Financial and taxation database
- Shocks and correlations
- Non extensive models
- Non Markovian models
- Agent impact and market response functions
- Bayesian belief networks
- Correlation functions of risk
- Microbial growth
- Non-Gaussian/power-law variants of Black-Scholes
options pricing model
20Network dynamics and topology
- Network topology and growth
- Universality and the mechanisms that determine
scaling exponents - Network dynamics and the interplay with network
geometry.
21Physics and socio-economic issues
- Generalised Lotka-Volterra models
- New models of social opinion
- Social impact models and effects of localisation
- Communication in societal, economic and financial
networks - Modelling of collective social intelligence