Risk assessment and decision support
Slope stability: Deterministic model
a) equilibrium of potential sliding mass Potential sliding mass = Force(resistance of sliding) / Force(pushing down as gravity) b) Factor of safety: F = (N*c)/Pd - stability number N: geometry, from table with ß -> d=D/H (where D is height from floor to start of slope and H is height of slope) -> N0 - shear strength: c [kPa] from ground investigations (shear strength related to depth) - unbalanced stress: Pd = gamma*H-gammaw*Hw (where gamma is the unit weight of soil/clay [kN/m^3] and gammaw is unit weight of water [kN/m^3], H is height of slope and Hw is different between linear slope to real one)
Slope stability: Probabilistic analysis for F
a) probability of failure pf: i) F<1 ii) m=ln(F)<0 b) addressing uncertainties - estimate intervals for parameters N and cu±sigma (caputres approximately 70% of values, cu,average=25kPa -> Vcu= 10%) c) calculation method - Monte Carlo simulation - point estimate method (PEM) - analytical solution: m=ln(F)<0 ? i): m = ln(F) = ln(N)+ln(cu)-ln(Pd) µm = µln(N)+µln(cu)-µln(Pd) ≈ ln(µN)+ln(µcu)-ln(µPd) sigmam^2 = sigmaln(N)^2+sigmaln(cu)^2-sigmaln(Pd)^2 ≈ VN^2+Vcu^2+VPd^2 ii) reliability index ß = µln(F)/sigmaln(F) ≈ lnµF/VF = (ln(µF)+ln(µc)-ln(µPd))/sqrt(VN^2+VC^2+VPd^2) iii) probability of sliding: Pf = standard normal distribution ø - ß
Slope stability: Triggering factors for landslide
a) uncertainty versus time - permanent uncertainties -> non-triggering - variable uncertainties -> triggering b) assumptions - geometry and shear strength -> permanent - external water -> variable
Road safety: Hydrogeological probability model
advective flow time - possibilities for transport through the unsaturated zone: f=infiltration capacity, volume of spill, area of spill, depth of groundwater, retention capacity, geological barriers - travel time to compliance boundary: f=distance to boundary, effective porosity, hydraulic conductivity, hydraulic gradient
Hazard identification and risk estimation
answering 3 question: what can happen?, how probable is it?, what are the consequences? - considering likelihood of an hazard having an effect
CBA: special to other analysis
anthropocentrism - focus on human wellbeing principle of consumer sovereignty and not experts opinion - focus on humans own opinions on their wellbeing utilitarianism - aggregation of wellbeing over individuals is acceptable and focus on consequences not on actions themselves
Consequences
human - disease, mortality,... ecological - number of affected individuals, genetic variability,... economic - damages to properties, delays, .... template - acute - restoration - long lasting effects
MCA: Ranking of alternatives
methods - multi-attributive utility methods (MAUM) -> includes uncertainty - value function (MAVT) - linear additive methods -> simplification of MAVT, assumes mutual independence of preferences - analytical hierarchy process (AHP) - outranking - non-compensatory methods -> dominance, threshold of performance, ranking of criteria
Event tree analysis (ETA)
- addresses question: what can this initial event lead to? - horizontal structure which proceeds in time from left to right -consists of: chance nodes (uncertain events) and terminal nodes (outcomes) - branches spread out from a chance node and represent possible events, each associated with a probability - steps 1) identification of a relevant initiating event that may lead to unwanted consequences 2) identification of the safety functions that are designed to deal with the initiating event 3) construction of ET 4) description of the resulting accident event sequences 5) calculation of probabilities/frequencies for the identified consequences 6) compilation and presentation of results from analysis
Uncertainty analysis: Monte Carlo methods
- all kinds of uncertainty measured with probability distribution -> probability used to quantify uncertainty by: 1) analogy: physical randomness includes uncertainty 2) axiomatic or normative approach 3) coherence of bets - result -> not only what could happen but also how likely it is to happen
Road safety: conclusion
- application of collection systems and hydraulic barriers: only motivated by societal risk-cost perspective at large water supplies adjacent to roads with high transport frequency - most cost-effective measure at many sites is increased preparedness at rescue party - decisions most sensitive to selection of failure criteria, time horizons and valuing of the groundwater resource
Hazard identification
- as many tools as possible - based on a conceptual model methods - checklists and brainstorming: identification of most hazards that lie in the operating experience among people involved -> disadvantage: do not encourage participants to expand thinking to new possibilities - structured brainstorming - hazard matrix: matrix of interactions between activities and components of the environment, helpful in identifying activities, construction depends mainly on checklists and brainstorming - hazard and operability analysis (HAZOP): structured, expert brainstorming, use of conceptual models and influence diagrams with guide words (as less of, more of, reverse flow,...) -> encourage group of experts to interrogate system and apply their expertise beyond their own experience, experts apply 'what if' question to each component of a system in structured manner - failure modes and effects analysis (FMEA): same primary objects as HAZOP procedures, detailed examination of causal relationships between elements in a system, calculation of risk priority number (FMECA)
Decision support
- basis for decision-making - comparing risk reducing options
CBA: mathematical description
- calculation of Net Present Value: NPVi = SUM_t=0_T_( (1/(1+r)^t) * (Bi(t) - Ci(t)) ) where B = benefits of alternative i, C = costs of alternative i, T = time and r = discount rate NPV > 0 -> social profitable the higher discount rate, the less profitable for society
CBA: conclusion
- compute the value of the objective function for all project options - likelihood that effects, that have not been possible to quantify, would change conclusion - carry out sensitivity analysis - carry out distributional analysis
Groundwater flow and contamination transport: transport processes - advection
- contaminant moves with the groundwater Darcy´s law Q=KA*(deltah/L) ; Q=K*A*i where Q~deltah; Q~1/L; Q~A; i=deltah/L - bulk motion of groundwater pore scale advection average groundwater velocity: L/T = v^ = K*i/ne where ne = effective porosity [%] - assuming uncertainty only in Kmean: v = K*i/ne Retention time: T= (L*ne)/K*i mass flux contaminants - specific flux (per unit area): F=v^*ne*C=qC - total flux: Ftot=FA=QC
CBA: uncertainty analysis
- data preferable given as intervals instead of point estimations - risk values are high and uncertain - Monte Carlo simulation
Risk analysis
- description of concerns - definition of system - " of circumstances - stating assumptions - identification of analysis decisions context, scope definition, problem identification - purpose - questions to answer - what information are required - involvement of stakeholders&their responsibilities - what is affected by decision and how - type of risk assessment - limitations (social, scientific, time, money) - problem of boundaries conceptual model - to understand system - important that all needed knowledge and expertise are available - may be used to evaluate risk options
Failure causes
- design failure (weakness) - manufacturing failure - use failure (ageing, misuse, mishandling) - external events affecting the system
Sustainable development
- ensuring that the use of resources and the environment today does not restrict their use by future generations - economic, ecological and social aspect -> sustainable solution only where these 3 meet
Groundwater flow and contamination transport: transport processes - diffusion and dispersion
- geological stratification and velocity conditions are key factors - contaminant spreads as it moves due to combined effects of diffusion and dispersion diffusion - caused by concentration gradients - dominant in low permeability media - generally negligible in high velocity environments - rate of diffusion described by diffusion coefficient dispersion - faster transport in large pores - some particles take longer routs than others - friction differences within pores
MCDA: Web-HIPRE
- hierarchical preference analysis in WWW - provides implementation on multi-attribute theory - structure the problem, prioritise, analyse the result - decision problem is visually structured into a value tree: overall objective, criteria and possible sub-criteria (attributes), options - define attribute ranges - value functions for attributes: direct, AHP, value function - weight elicitation: direct, SWING, SMART, AHP - results and sensitivity analysis
Groundwater flow and contamination transport: groundwater pollution source
- industry - urban runoff - sanitary landfill - septic system - illegal dump - agricultural chemicals - spill - municipal sewage treatment plant
Groundwater flow and contamination transport: key question
- is time t sufficient? - proper actions needed? - is concentration at receptor (well) acceptable? - what amount of contaminants will reach the receptor? what is contaminant load [kg/a]
Decision analysis
- methods to support decision under uncertainty - components of a decision model: decision options (alternative actions), relevant outcomes of these options, probabilities of the outcomes (if analysis is probabilistic), decision rule (aims of the decision-maker) - Keeney's definition: "formalisation of common sense for decision problems which are too complex for informal use or common sense" models - from data collection to conceptual, parameter, probability and consequence models - problem specific probability model results out of conceptual and parameter model - all contribute to decision model: failure criteria and risk reduction alternatives
CBA
- performed on societal level - prioritising of measures needed since limited resources - relevant for sustainability evaluation - analysis of distributional effects necessary - common language is key for efficient solution when several interested parties involved in decision-making - estimation of changes in public welfare in terms of money -> expression of changes in public welfare (positive and negative expressed in monetary terms) -> comparison possible since all impacts are measured in the same unit (money) as far as possible basic criterion - sum of all benefits for all affected parties must be larger than the sum of all costs for all parties - benefit = risk reduction
Reasons for risk management (RM)
- preventive rather than reactive - control, prevent and reduce loss of life, disease, damages to properties,... (IEC) - facilitation of rational decision making - distinction between greater and lesser risks - creation of opportunities - increase of awareness and knowledge regarding risk issues - support of risk communication multi-disciplinary - system analysis - probability and statistics - engineering - physical, chemical and biological sciences - health and social sciences - decision analysis - human factors and management science
Road safety: consequences
- remedial costs in unsaturated and saturated zone - costs of temporary or permanent replacement of water supply -....
Risk evaluation
- risk tolerable or not? -> tolerability criteria - analysis of risk-reducing alternatives/options - decision criteria -> cost-effectiviness, cost-benefit, sustainability
Project risk
- unexpected delays - liability - worker accidents - accidents to 3rd party - political risks -....
Estimation of risk reduction
- use of fault tree method and cost - effectiveness analysis for evaluating risk-reduction measures - method 1) hazard identification 2) analysis of system -> FT 3) evaluation of risk levels 4) identification of possible risk-reducing measures 5) analysis of risk-reducing options: cost-effectiveness and fault tree 6) evaluation of results 7) input to decision-makers
Ecosystem goods and services
- whenever humans derive wellbeing through something in nature (direct or indirect) -> natural capital categorising: - supporting e.g. biochemical cycles - provisioning e.g. food - regulating e.g. climate regulation - cultural e.g. recreation in CBA - no deltaPS or delta CS - economic value of non-market goods by special evaluation methods a) revealed preference (RP) methods: travel cost method, property value method (hedonic method), defensive expenditure method, production function method b) stated preference (SP) methods: contingent valuation method, choice experiments
Decision which risk needs to be reduced
-> risk matrix the higher probability and consequence the less acceptable is risk
Fault tree analysis (FTA)
-addresses the question: how can this event occur? - structured process that identifies potential causes of system failure - top, intermediate and basic events - illustrates/models the interactions between events using logic gates: OR-gate (series system, at least one of the input events must occur), AND-gate (parallel system, all input events must occur) performance - deductive technique: what are the reasons for this event? - binary analysis [0,1] - input data = probability for the basic event - steps: 1) definition of problem and the boundary conditions 2) construction of FT 3) qualitative analysis of FT 4) quantitative analysis of FT interrelationships between a critical even and the causes of this event - possible causes: environmental conditions, human errors, normal events and component failures - possible results: list of combinations of events that may cause system failure, probability of system failure (top event) dynamic time behaviour
CBA: to do
1) calculate annual risk reduction = benefit 2) choose time horizon 3) choose discount rate 4) estimate costs for implementation and maintenance for each alternative and each year during time horizon 5) calculate NPV for each alternative 6) perform uncertainty analysis
Process RM (according to IEC)
1) risk analysis - scope definition - hazard identification -> start update loop - risk estimation 2) risk evaluation - risk tolerability decisions - analysis of options 3) risk reduction/control - decision making -> end update loop - implementation - monitoring -> continuous updating -> iterative process obtain info and knowledge about risk -> causal and consequence analysis (qualitative and quantitative) -> uncertainty analysis
Probabilistic Risk Analysis: Bayesian methodology - errors
1st kind - mild error - P(D/F^c) -> false positive -> detection where no unwanted event 2nd kind - significant - P(D^c/F) -> false negative -> no detection where unwanted event occurs
CBA: distributional analysis
3 common approaches - ignore distributional aspects: hard to defend, but often used - compute how costs and benefits are distributed among different (income) groups - use different weights for the NBs of different groups
Decision-making
= choice of one risk-reducing option among a number of alternatives process of making a choice - define decision problem - assess problem - collect and verify information - identify alternatives - anticipating effects of decisions - make choice by using all available info - evaluate decisions - communicate decisions kind of stakeholders - those exposed to risks - those benefiting from risk generating activities - decision-makers
Groundwater flow and contamination transport: transport processes - sorption
= retardation -> use of barriers - contamination movement is slowed down due to reaction with other solutes and the geological media due to adsorption and absorption processes retardation factor R=v(groundwater)/v(plume) - slowing of contaminant migration distribution coefficient Kd - concept: sorption of contaminants to geological media or organic content of media R = v^/vc^ = 1+(rohb/teta)*Kd = 1+(sorbed conc./mobile conc.) where rohb = bulk density of porous media [M/L^3], teta = moisture content or porosity for saturated media [-] C*=Kd * C where C = concentration of solute in solution in equilibrium with mass of solute sorbed onto the solid [M/L^3] and C* = mass of solute sorbed by dry unit of weight of solid [M/M]
Methods in decision support
CBA = cost-benefit-analysis - NPV - risk reduction = benefit Conceptual economic model - costs of risk decrease with increase in risk-reducing alternative while risk reduction costs increase - optimal risk is where total cost (combination of former ones) is lowest MCA = multi-criteria-analysis - identification of sustainable alternatives and ranking based on multi-criteria
Cost effectiveness of risk-reduction measures
CERj = cost-effectiviness-rato = Cj/Ej where the cost is Cj = SUM_t=1 to T_(Cjt/(1+r)^(t-1)), E = effectiveness, r = discount rate, t = year analysis - tool to identify what alternative achieves a specified target to the lowest cost
Process RM (according to Burgman)
Context -> Formulate problem/engage stakeholders -> Set assumptions -> Develop/ update conceptual models -> Assess hazards/ assessments for range of scenarios -> Rank/ calculate/ project risk -> Analyse sensitivities -> Decide -> Monitor (-> Formulate problem or -> Set assumptions or -> Develop/ update conceptual model) -> continuous updating -> iterative processR>isk
Event and process orientated models
Event orientated - conditions under which event occurs Process orientated - aim to describe process
Slope stability: Nominal levels
HHW = highest high water = highest observed level (M)HW = High water = Mean of annual high water MW = Mean water = Mean water level with time (M)LW = Low water = Mean of annual low water LLW = Lowest low water = Lowest observed level uncertainty of water level - low water level -> risk of sliding - input for estimation of annuaö probability of sliding = LW ± [uncertainty of LW]
Combination of FTA and ETA
Hazard identification -> Causal modelling (FT) -> Probability of failure -> Consequence modelling (ET) -> Estimation of risk
Purpose of RADS
Identification, evaluation and design of cost-effective and sustainable actions for controlling risks and increase safety in projects
Case 1: Road safety and risk control
Key questions - will accident with contaminant release occur? - will water quality be degraded in case of release? -> what are possible consequences? - how effective are possible protection measures? - what are protection costs? R=P(f)*C(f)=probability of failure * consequence in case of failure Risk reduction - preventive measures for reduction of probability of failure: e.g. increased road safety, increased technical reliability of tanks, collection systems - measures towards reduction of consequences: e.g. alternative water supplies, hydraulic control, increased preparedness - combination of both! Decision models - data collection -> conceptual hydrogeological model, parameter model, accident and spill probability model, consequence model - CHM&PM -> Hydrogeological probability model - HPM&CM&ASPM -> DM including failure criteria as travel time between accident location and compliance boundary and maximum allowable concentration at a compliance boundary
MCA
Multi-Criteria-Analysis - evaluation and choice of alternatives based on multiple criteria using systematic analysis - transparency -> tool for communication between parties - scores assigned to performance criteria can deviate substantially between different decision-makers -> analysis should be performed by team - simple, but results may be difficult to interpret when no documentation is done scores and weights - used to make sustainable choice - calculation of resulting weighted performance score: Final score = SUM_i=1_N_(Wi*Si) where S = score for criteria i and W = weight of criteria i
Bayesian - terminology
P(F) = prior probability P(F/D) & P(F/D^c) = pre-posterior (before test) and posterior (after test) P(D) & P(D^c) = prior predictive probabilities
Risk and uncertainty
P_F = estimation of the proportion of time, the system cannot meet safety targets = (1/lamda) / ((1/lamda)+(1/µ)) where 1/lamda = time to failure and 1/µ = duration of failure = MDT / (MTTF + MDT) where MDT = mean downtime and MTTF = mean time to failure = lamda / (lamda+µ) R = SUM(C*P) - C modelled by Beta distribution - lamda and µ modelled by gamma distributions methods used to bring all models intro on - Bayesian approach - Monte Carlo simulations
Risk priority
R = P*C R = S*O*D where S = severity (of effect of failure); O = occurrence (likelihood that a particular cause leads to failure mode during a specified time frame) D = detection (likelihood that the current controls will detect the cause of the failure mode and preventing it from occurring
Bow_Tiw_diagram
With time: Causes -> safety barriers -> initial event -> safety barriers -> possible consequences
CBA: risk valuation
as less as reasonable practicable (ALARP) - risk magnitude small - tolerable only if risk reduction is impracticable or if cost exceeds benefits 2 approaches - no general rule for relationship between them 1) ex post estimate WTP for avoiding a consequence as if it occurs with certainty - WTP_ex_post = max WTP for a project that would avoid the occurrence of the adverse event A - multiply WTP_ex_post with p that A occurs 2) ex ante estimate WTP avoiding a risky situation - both probability and consequences of A taken into account - advantages: take into account that individuals might have different opinions towards risk (depending on type of risk, size of risk in initial situation, size of risk change), consistent with principle of consumer sovereignty - disadvantage: risk perception issues: objective vs subjective
MCA: decision dilemma
balance between - different choices = different costs - impacts on human health - public opinions - ecological impacts
Monte Carlo: Important elements
calibration - adjusting model parameters, structures and assumptions to fit available data and intuition sensitivity analysis - calculating the magnitude and rank the order of responses of consequences as a function of model parameters, assumptions and model structures - identify which variables have greatest impact on the outcome validation - comparing independent fields observations with predictions -> testing ideas
Probability - chance, belief, tendency, confidence
chance - frequency of a given outcome - among all possible outcomes of a random process or within a given time frame belief - degree to which a proposition is judged to be true - reported on an interval (0,1) or %-scale -> analogy with chance tendency - physical properties of a system that result in a stable long-win frequencies with repeated trials confidence - degree to which we are sure that an estimate lies within some distance of the true value
Risk definition
combination of the probability and the consequence of an undesired/hazardous event: R=P*C 3 questions: - what can happen? - how likely is it to happen? - what are the consequences? affection of individuals (high probability, "low" consequences), groups and society (low probability, high consequences) Risk reduction - preventive measures for reduction of probability of failure - measures towards reduction of consequences - combination of both!
Groundwater flow and contamination transport: decay
contaminants degrades with time - factors affecting half-lives e.g. temp.(high->shorter lt), moisture content (low->shorter), organic content (low->shorter), adsorption (decreased->shorter) first order decay Cc(t) = c0*e^(lamda*t) = contaminant concentration at time t [M/L^3] where c0 = contaminant concentration at t=0 [M/L^3] and lamda = decay constant
Risk based decision analysis
cost-benefit -> CBA multi-criteria -> MCA
Risk assessment as part of RM
covers risk identification, analysis and evaluation
Perception of risk
differs between those who are exposed, those who benefit from risk generating activities and decision makers -> affecting factors - catastrophic potential - familiarity - uncertainty - individual (one person - e.g. by drinking water quality) or social (e.g. epidemiological) -> health and safety - controllability - voluntariness -> forms of risk perception risk averse - rather smaller reward with greater certainty than larger reward with less certainty risk seeker risk neutral
Groundwater flow and contamination transport: governing equation
dispersion - advection - sorption + decay
Monte Carlo: stochastic model building
distribution of model prediction - update again and again - probability distributions for any factor that has inherent uncertainty as input: beta, norma, lognormal, uniform, triangular,... f(x), f(y), f(z) -> f(x,y,z) sample approach 1) develop deterministic model 2) add stochastic elements to represent uncertainties 3) add assumptions about dependencies ->!crucial to take into account 4) use stochastic models to estimate statistical distribution of result -> simulation: values samples, at random from input probability distributions, set of sample = iteration 5) compare result with reality and update model
Process RM (according to ISO)
equivalent to IEC but including monitoring/review and communication/consultation in each step (establishment of context, risk identification, risk analysis, risk evaluation, risk treatment)
Statistical inference: Hypothesis rejection: errors
error of 1st kind: direct rejection at e.g 90 % -> unlikely but possible that something is wrong error of 2nd kind no rejection at any level -> power of the test is not strong enough to detect error
Road safety: Accident and spill probability models
f=annual frequency of petroleum transports, road location, expected annual frequency of road accidents, expected number of vehicles involved in each accident , speed limit, length of road section (the longer, the higher the probability), condition adjacent to road (rocks, trees, ..) probability of an accident for road stretch = P0 P0=N*Q*L365*F*10^-6 with N=average number of petroleum transports, Q=accident index (number of accidents per million transport kilometres), L=road length (km), F=number of vehicles per accident probability of a spill in case of accident = PU - influenced by road setting, speed limit and type of road probability of accident and release of contaminant P=P0*PU
calculation of risk
fault tree analysis & event tree analysis - event orientated, or logical models describe conditions under which events occur -are composed of conditions and logical terms, usually with binary space uncertainty modelling -> Monte Carlo
CBA: Economic valuation
for firms' part of economy - changes in Producer Surplus (PS) ≈ profits PS = TR - TVC where TR = Total Revenues and TVC = Total Variable Costs - changes in producers wellbeing = deltaPS because of deltap - max profits - if market price is p^0 -> (TR)=p^0*q^0 - competitive market: firms' marginal cost curve = supply curve -> MC = dTVC / dq for consumers' part of economy - changes in Consumers Surplus (CS) ≈ difference between maximum amount an individual is willing to pay for a good (WTP) and what it actually has to pay = total WTP - expenditure where expenditure = p^0*q^0 - max utility - market price p^0 increases -> demand q^0 decreases - demand curve gives info on max WTP for consuming one additional unit; total WTP is whole area below demand curve both need to be included in CBA - perfect competitive market where price at a level to reach highest possible demand -> boarder of expenditure to CS
MCDA: Stakeholder involvement
framework that permits stakeholders to structure their view about pros and cons of different options typically by means of workshops or surveys some applications attempt to represent the value judgement of one singe decision maker - stakeholder values are then considered as an attribute in the model e.g. what would public think about model
CBA: structure
important characteristics - problem formulation: precise project alternatives, including reference alternative (null alternative) - analysis ex_ante or ex_post - definition of social desirability: Hicks-Kaldor criterion = compensation criterion - systematic book-keeping of benefits and costs on a societal level - benefits and costs expressed by using monetary measures of changes in wellbeing in general - compute NPV of each project option (i) in relation to the reference (null-) alternative as the discounting sum of benefits (B) minus costs (C) for all affected in all time periods, where r is the discount rate NPVi = SUM_t=0_T_( (1/(1+r)^t) * (Bi(t) - Ci(t)) ) 4 main steps 1) define objective function - what are the project options (incl. 0-alternative) to be analysed 2) identify costs and benefits - which costs and benefits are relevant? 3) quantify costs and benefits - how to monetise?what valuation methods? what cannot be quantified? 4) compute net benefits - what do the results say? - what are the uncertainties?
Case 2: Slope stability
key questions - what can happen, how likely is it and what are the consequences in case it happens
Decision alternatives
no protective measure/ no action - severe consequences and/or high probability of the unwanted event to happen -> risks are unacceptable different protective alternatives (risk-reducing measures) - tolerable consequences and/or low probability of the unwanted event to happen -> risks are tolerable or as low as reasonable practicable
MCA: objectives, criteria, attributes
objectives - something that one desires to achieve e.g. the most sustainable solution, to find the best job,... criteria and attributes - used to determine how well the objective is met - attributes sometime described as a measurable criterion
Case 3: Groundwater flow and contamination transport
principe: Source - pathway - receptor source - contaminant - source barrier transport pathway - pathway barriers: pipings, surface water, groundwater Exposure at receptor - receptor barrier
CBA: disadvantages
problems - data demanded is extensive -> too few data -> generalising -> only reasonable in large projects - changes related to environmental impacts on society difficult to monetise
Structure of decision-making -> role of risk analysis and decision analysis
provide decision support, but no decision
CBA: advantages
provides - structure - transparency - help to identify and focus on factors that may otherwise be left out -qualitative estimations
Risk assessment of technical systems
risk analysis, - scope definition - hazard identification - risk estimation risk evaluation - risk tolerability decisions - analysis of options risk reduction/control - decision-making - implementation - monitoring
Probability - dimensions
statistical frequency: - long-term probability with some confidence limit - e.g. throwing a dice degree of belief (Bayesian) - probability is known (or not possible known) - follows rules of probability update as new information becomes available
MCDA tools
used to assess values judgement of individual decision makers or multiple stakeholders - group problems -> process of quantifying stakeholders preferences -> very intense