Business process simulation how to get value out of it (no magic 2013)
Persentation to NoMagic World 2013
Business Process Simulation: How to get value out of itDenis Gagné,www.BusinessProcessIncubator.comChair BPMN MIWG at OMGBPMN 2.0 FTF Member at OMGBPMN 2.1 RTF Member at OMGCMMN Submission at OMGChair BPSWG at WfMCXPDL Co-Editor at WfMC AbstractBusiness Process Simulation can be an effective tool when looking for optimal performancefrom a Business Process Model. Although considered quite relevant and applicable in thecontext of Business Process Management (BPM), Business Process Simulation is not currentpractice for -and even seldom used by- Business Analyst in the course of process analysis.In this presentation we will explore why this may be the case and will discuss how to useBusiness Process Simulation efficiently while identifying some of the pitfalls along the way.Various Business Process Simulation approaches, their benefits and applicability will beintroduced. The session will conclude with a quick overview of a new Business ProcessSimulation standard that is emerging within the industry. Poor Performing Processes May lead to: Delays Back log Refund Claims Angry customers Lost of goodwill (Mission Critical) Lost of lives (Life Critical) Gain Insight: Thoroughly analyse business process in a safe isolated environment prior to Deploying Simulation for Process Analysis Provides a priori Insight Can be Effective Process Analysis tool for: Alternative Evaluation Decision Support Performance Prediction Optimization Benefits of Simulation Advantages of simulation over testing on the real world include: Lower relative cost of business transformation explorations Speed of validation of potential scenarios No disturbance to current operations Simulation for Business Processes Visual Depiction (Visualization & Animation) User is presented with a (sometime interactive) animated depiction of the business process model Numeric Simulation (Discreet Events) User asked to provide values for input and decision parameters of a simulated business model Role Play (Serious Gaming) User asked to take actions and make decisions within a simulated business environment Types of Process Analysisusing Simulation Structural Analysis The structural aspects (configuration) of a process model Usually Statistical Analysis (using static methods) Capacity Analysis The capacity aspects of a process model Usually Dynamic Analysis (using discreet simulation methods) When is Numeric Simulationmost Appropriate Capacity analysis of processes that potentially are Highly Variable Variability makes outcomes difficult if not impossible to predict Interdependent Changes in one process affect other processes Complex Complex structure or complex behavior Capacity Constraints Hard resources constraints (as independent variables) When is Numeric SimulationLess (Not) Appropriate When an expedited analysis indicates a negligible problem When there is little or no variability or uncertainty When the consequences of poor estimates are acceptable When the cost of intervention is less than the cost of the analysis experiment Process Simulation Other Uses Training & Learning Although very popular in support of operations, limited use in other Business disciplines Persuasion & Selling Simulating results of a proposed solution Cause and Effect simulation Verification & Validation Validation: Are we doing the right “thing”? Verification: Are we doing “it” right? Process Simulation not yetCommon Practice: Why? Potential Reasons: Availability Limitation of existing BPMS Tooling Lack of Training or Expertise Lack of Standards Optimization Selection of a best scenario (with regard to some criteria) from some set of available alternatives Almost impossible without tool support Sub optimization caveat Optimizing the outcome for a subsystem will in general not optimize the outcome for the system as a whole. Process Improvement ProjectBest Practices Defining Success “Can’t get there if you do not know where you are going” Why are we conducting this project and what are the objectives Stakeholder Analysis “When it comes to assessing success your own opinion while interesting is irrelevant” How do your stakeholders define success While it is obvious that satisfying the most important stakeholder is necessary, it is rarely sufficient. Do not ignore other stakeholders Process Improvement Projectusing Simulation Get the Goal Right Clearly define the goal or problem to be investigated using simulation Clearly state the objectives of the simulation investigation Match Expertise to Desired Experimentation Different levels of Investigation Complexity Get the Model Right Model Granularity Model Parameterization Clearly Define the Goal Intentions Examples Reduce headcounts or expenses Improve process predictability or reliability Increase throughput Increase output Ensure SLA Design the Experiment Independent vs dependent variables Same process model under different parameterisations Different process models under same parameterization Number of distinct model settings to be run The experiment should provide insight The experiment should help inform a decision The experiment should be in response to clearly defined objectives that are relevant to a decision Clearly Define the Objectives Provide SMART Objectives Specific Usually answer the five “W” questions Measurable Aiming for quantifiable, concrete results Achievable While an attainable goal may stretch a team in order to achieve it, the goal is not extreme Relevant To your boss, your organization, your stakeholders Time Bound Within a time frame, with a target date Be mindful of the Optimization Conundrum Optimization Conundrum Quality Time Satisfaction Lean Cost Expertise vs Experimentation Expert Verify Process Structure and logic OptimizationProcess Modeling Novice Learning via Quantitative Experimentations Analysis Novice Expert Simulation Model Granularity Pick the right level of process model abstraction e.g. What is an atomic task For example a certain level of details may suitable to compare relative throughput of alternative process designs while not be detailed enough to provide reliable prediction of actual throughput Model Input Parameterization Setting Input parameters for process model elements to reflect external stimulation e.g. Arrival Patterns Opportunity to introduce event variability into the process model Select Candidate Probability Assess Fidelity Can easily be the cause of misleading results “Garbage in garbage out” Select Candidate Probability Based on the external observed behavior Is it Constant or Random Select a distribution that best captures characteristics, observations, or available data Some distribution are better fits to specific situations (e.g. Poisson for mutually independent arrivals) Using available historical or event log data as reference may require data cleansing e.g. minimum task time =8 mins Mode task time = 32 mins Maximum task time = 9.5 hours May not notice that maximum task time includes an 8 hour off shift Assess Fidelity Check how well your input parameterization reflect the observed behavior Model behave as desired or expected, or Model behavior reflects “As Is” situation Carry out Sensitivity Testing Determine how sensitive your model is to different input parameters Check sensitivity in magnitude (e.g. mean) and variability (e.g. range) Simulation is often a process of discovery Examine output results Unexpected result are not necessarily a problem Primary reason for your simulation experimentation Need to find an explanation Will provide enlightenment of actual process behavior vs assumed process behavior Unexplainable results are a problem When Examining Results When randomness is introduced replications should be used Replication = same scenario but with different sequences of random variables e.g. repeated coin toss Warm up periods may be required Reflect the notion of work in progress (WIP) Time during which results are either not collected, or which can be separated off from the main results collection period e.g. A bank (opens empty and idle each day) model does not require warm-up (and indeed should not have warm-up). Common examples of situations requiring warm- up are manufacturing in general, hospital emergency rooms, 24-hour telephone exchanges, etc Demo Randomness and likelihood Business Process Simulation Working Group BPSWG www.BPSim.org Why BPSim Encourage wider adoption of simulation within BPM community through a standards led approach Process simulation is a valuable technique to support process design, reduce risk of change and improve efficiency in the organisation Provide a framework for the specification of simulation scenario data and results as a firm foundation for implementation Open interchange of simulation scenario data between modeling tool, simulator, results analysis/presentation tool BPSim Scope Complements existing process modeling standards “Not Reinvent the Wheel” BPSim Element Parameters Each element parameter of a scenario references a specific element of a process within the business process model Each element of the business process model may be parameterized with zero or multiple element parameters Perspectives TimeParameters ControlParameters P ResourceParameters CostParameters InstanceParameters PriorityParameters Demo Business Process SimulationBest Practices The Right Model for the Right Goal Align Modeling Objectives with Simulation Objectives Abstraction Fidelity Validity (soundness and completeness) The Right Answer to the Right Question Make sure to instrument your business process model with parameters that are actual indicators (influencers) of what you wish to explore The Right Expert for the Right Task Although conceptually simple to grasp, successfully (meaningfully) using numerical simulation for business modeling still requires some expertise (Advanced Mathematical Skills) Business Process SimulationCaveats Unrealistic User Expectations Simple Press-Button Simulation Deterministic behavior assumptions A Business Process Model is a Simulation Model (not necessarily) Their goals (purposes) may be misaligned Be Mindful of Misleading Results (Garbage in Garbage Out) A simulation model that is fidel &valid with uncharacteristic data can lead to incorrect conclusions or predictions, Negative Training, … Sub-Optimization Partial or sub-model optimization can lead you astray Discussions & Questions www.BPSim.org email@example.com
Leave a Comment
You must be logged in to post a comment.