Dynamic project management using simulations
One of the toughest jobs of a Program Manager is to make decisions to correct the course of a program and ensure that all the projects in the program are following the defined flight path. More often than not these decisions are based on 30% data, 30% group think and 40% intuition that is really past experience!(CGN Analysis) In the engineering world we have been using simulations (computer programs that mimic or model behavior of physical systems) for years to predict and explore system capabilities and limitations. These simulations have saved millions of dollars and in many cases even lives. Applying the engineering simulation techniques to program plans can have a similar impact on the program. It gives the program manager the ability to model and view the impact of changing any given constraint on a project on the entire program and all related projects. The simulation can also give an insight into the different level of interactions between projects and the intensity of these interactions. Now without actually implementing the new course or even changing the project plans, the program manager can view the new flight path, cost, schedule and quality implications.
This paper provides an overview of how simulations assist Project managers in decision-making and program analysis. This paper also touches upon some leading edge simulation tools and techniques for business scenario analysis and project management that can substantially increase the overall probability of success by creating a program roadmap, managing risks and making business decisions based on predictive analytics.
Most people in the business world expect their Projects to be completed on time, meet quality objectives, and fall within the allocated budget. To most managers, time and money are critical factors to be carefully considered in any decision. These various factors make Project Management extremely challenging for organizations. Today’s business leaders recognize that advanced decision making capabilities using simulation techniques can directly improve their organization’s performance. To do so, however, a Project has to be meticulously planned, effectively implemented and professionally managed to optimally achieve the objectives of time, cost and performance.
Some of the key findings from surveys conducted by different organizations are:
- Of 3,874 projects, 47% were troubled, troubled and recovered, or troubled and failed in early 2006. (PMI, 2006a, p16.)
- Cost overruns are as high as 189% over the planned budget and 224% over the planned time estimates. (Source : Standish Group 1994 CHAOS Report, as reported in SofwareMag, Feb/Mar 2001, retrieved from http://www.softwaremag.com/archive/2001feb/CollaborativeMgt.html on 07.12.06
- In a study conducted by BIA, Toronto, Canada 73% of organizations that eliminated their PMO said the primary reason was “lack of executive sponsorship”. (PMI, 2006b, p12.).
Many Projects experience schedule slippages and cost overruns are due to reasons that are both varied and numerous. As there are literally thousands of variables, both human and non-human, in a Project, the only way to combat the enormous uncertainties that would be easily generated out of such large varieties is to develop sound systems to manage them.
Dynamic Project Management using simulation is a new trend in streamlining the management of Projects. Complex plans can be analyzed by a computer and applied in many critical areas. This practice enables the Project manager to address issues by simulating their impact on the project before they become problems. This allows Projects of any size and complexity not only to be planned but also modeled to answer ‘what if?’ questions.
How Simulations help in Project Decision –Making
Simulation modeling is useful in Project decision-making by looking at problems as a whole and articulating the complete set of relationships, interactions, and uncertainties. A business simulation model can predict what may happen in the physical system and take necessary steps beforehand, so that the physical system can maintain equilibrium despite various disturbances from both within and outside. Project planning and simulated models can be exploited to provide real time information, as well as linking to risk-analysis, time recording, costing, estimating and other aspects of Project control. Decision makers can virtually explore future direction opportunities in a safe, cost-effective forum without having to experiment on a real system. This helps in moving from traditional “rear-view” analysis to the ability to see the predictable future, control it and take actions today to attain tomorrow’s goals. Benefits include providing critical business insight into the project, reducing time, resources and risks; increasing the quality, utility and supportability of systems; and enabling integrated product and process development.
In all instances, Project work is a dynamic system. However meticulously the model is drawn, it will soon outlive its purpose unless it is continuously updated by obtaining feedback from real life. It is advisable not to build the whole model at once but rather take a modular approach where components are added a little at a time while continuously testing and validating the outputs at each stage. Simulations are effective when you cannot predict what the effects of change will be or how the proposed line might work, when you need to further understand a line or factory, you need to test a new or proposed system before investing in the equipment, or you have to refine an existing process. For example an engineering program has a complex set of varieties. The varieties are in terms of technology, equipment, materials and people, work culture, and ethics. These varieties can be combined to create a predictive model that can be used to simulate forecasts of how the program would behave over time. The intelligence of the model would come from careful definition of the business conditions, the behavior rules of the independent varieties, the dynamics of the interactions between these varieties, the frequency of such interactions and a whole slew of other patterns that must be programmed into the model. Dynamic simulation also helps visualize multiple organizational processes simultaneously, resulting in a quantifiable solution set matrixed against organizational value drivers. Consequently, organizations can derive benefit from improved strategies, policies and risk-mitigation plans; allowing the management to make proactive and intelligent decisions. (Exhibit 1) explains the advantages and issues in using simulations
Simulation Modeling v/s Spreadsheet Analysis
Use of spreadsheets is very common in analyzing data and creating “what- if” scenarios. Although spreadsheets are very useful and a powerful tool for analyzing data and creating forecasts, use of simulations provides a distinct advantage when trying to model and predict the behavior of complex programs and business scenarios. Also once the system is correctly modeled the power to look into the future can only be provided by simulation. Following are some of the key advantages of using simulation techniques over spreadsheets.
- Spreadsheets use averages to represent schedules, activity times, and resource availability
- Spreadsheets cannot accurately reflect the randomness and interdependence present in reality with resources and other system elements
- Spreadsheets are static and can only provide a quantitative result for one moment in of time, where as simulation replicates the dynamic business reality (the power to look into the future)
- Spreadsheet does not have the flexibility of animation, where as simulation can clearly provide visibility to the effects of a local change on a global system
Examples of how simulated data can be useful to the decision process
Consider a manufacturing firm that is contemplating building a large extension onto one of its plants. Company leaders are not sure whether the potential gain in productivity would justify the various costs associated with the expansion. It certainly would not be cost-effective to build the extension and incorporate it into their business plans and processes, only to remove it later if it does not work out. However, a careful simulation study could shed some light on the question by simulating the operation of the plant, as it currently exists and as it would be if the plant were expanded. Along with the more apparent issues, there are also less visible – but extremely important – concerns to be identified and factored into the overall equation. How would the workforce be structured and would be the make-up of the team? What would be the optimum mix of in-house talent and outsourced resources? What types of communication strategies would need to be in place? What types of system dynamics would need to be taken into account? What would be the impact of using various options? What kinds of risks are associated with even the best of scenarios? Traditional process analysis or project planning does not address these types of behavioral issues, and the usual way to resolve them would be to run the actual process, create detailed plans and observe the consequences-which would prove to be a costly experiment if it failed. In this case, simulation can assist in decision making in the following areas:
Simulations can be of major help in pinning down the requirements early in the product lifecycle, particularly for examining temporal behavior. This method is more accurate than costing models. Simulation can also account for dependencies between tasks, finite capacity resources, and delays resulting in rework loops.
Simulations allow managers to make more accurate predictions about the schedule, resources and the accumulated costs associated with a program or a project by creating a robust project design. This starts with establishing a framework for modeling and scenario analyses of the program plan. The framework is put into place prior to the detailed planning and WBS sessions.
The project template is designed with particular emphasis on the Products or Deliverables, the Goals or Tasks as they relate to the Products, the complex information and re-work dependencies between Goals, the first definition of project phases and any high level dependencies that may exist between them, and an understanding of the work effort that is required to achieve the goals that are defined in the model. This first simulation result is not necessarily accurate or optimum, but points out areas of design that need clarification or change to correctly reflect the real-life conditions and relationships.
The next phase begins with the creation of specific baseline models and simulations for the program and each of the sub projects leveraging the templates and specifications from the first phase. Also, a strategy is considered for the scope of scenarios once these sub projects are combined into a common program model. The most compelling of the scenarios are then explored and multiple “what-if” questions are asked, modeled, and simulated. The forecasts show the bottom line cost, duration, and risk of the overall program and the sub projects given these different scenarios.
The biggest advantage is that very early in the planning phase the executive team can get a more definitive understanding of the inter-Goal or inter-Phase dependencies, an understanding of the possible organizational revisions or flexibility that is possible, awareness of the capabilities and talents of the teams that are involved, knowledge of changes or requirements in Goal priorities that are possible and a clear understanding of the work process and the ability to modify it.
Simulation sensitizes managers to the consequences of instabilities resulting from badly designed organizational processes. Processes can be identified, improved and rolled-out to staff with few surprises.
By applying simulation modeling to static process information flows, such as Value Stream Maps, it is possible to analyze potential new scenarios in a virtual world. Process Simulation can help make intelligent change decisions when trying to understand the impact on the organization from restructuring and outsourcing strategies by simulating work process changes or resource pooling.
Once a static representation of the process is ‘brought to life’ in the virtual world, individual activities, resources, times, bottlenecks and other independent variables can each be made to behave and interact as desired so as to create anything from a gradual to a radical change in the original process. The predetermined performance indicators can be continuously calculated to evaluate the behavior of proposed changes to the process.
There are two primary powerful advantages of using simulation for process improvement analysis. First, these virtual process models present the opportunity to evaluate the pros and cons of multiple options before deciding on a final process to evolve to. Further, more than one option can be connected in series as a way to understand a phased approach for achieving the desired final state. This leads to the second advantage of simulation which is the dynamic capability of running the model by simulating time itself. The ability to study the impact of the independent variables on the process over a period of time can provide very useful insights with regards to potential bottlenecks in the future or other unforeseen events our outcomes.
Models can be developed to examine the potentially complex interactions of different variables and the impact of different risks on the overall system. Mitigation strategies can be developed that will lessen the impact of unanticipated and disruptive events. It also helps the management to focus on the most critical risk areas rather than all the identified potential risks.
Steps in Program Analysis using Simulations
The first step is to Identify the key business conditions and the organizational breakdown for the assignment of goals. Also identify the product breakdown and determine the proper level of work breakdown that will achieve the desired forecast accuracy for the project. Typical data and information that is required to achieve rapid progress in this phase are: Available resources (numbers, talents, etc) and their possible locations, hourly rates, work schedules, reporting relationships, the basic work process and the vision for product and work deliverables. This is the phase of the process where the background of the initiative is defined - participants discuss the organizational makeup, vision of the project and high level objectives. Key financial accounts are also identified and an activity-based costing model can be developed if required.
Next, a Model that provides a hypothesis based solution is formulated. This is the phase of the process where the specific initiative roadmap is built with particular emphasis on the Products or Deliverables, the Goals or Tasks as they relate to the Products, the basic dependencies between Goals, the first definition of project phases, any high level dependencies that may exist between them, and an understanding of the work effort that is required to achieve the goals. It starts with construction of a program roadmap at the top level and then develops detail in the area being studied. The key is to collect data on the system of interest, though only enough to capture the essence of the targeted system, and use it to estimate input parameters and to obtain probability distribution for the random variables used in the model.
After Identification and Modeling, the next step in the process is to visualize multiple project scenarios simultaneously through Simulation. Factors involved in this step include validation and verification, sensitivity or ‘what-if’ analysis, optimization, and uncertainty analyses. This is the part of the process where the project team examines the initial results, outline areas for improvement, decide on changes and perform multiple simulations that result in an acceptable and likely outcome for the project or program.
Information required for this phase includes more definitive understanding of the inter-Goal or inter-Phase dependencies, an understanding of the possible organizational revisions or flexibility that is possible, awareness of the capabilities and talents of the teams that are involved, knowledge of changes or requirements in Goal priorities that are possible and a clear understanding of the work process and the ability to modify it.
The final step is Documenting the assumptions that went into the simulation and Implementing the results of the study in the actual business situation. This is the final phase of the process where the results of the multiple scenarios are analyzed and where one or more that meets the needs of the organization is selected. The key to this selection is the clear understanding of the actions that are necessary by all to achieve the end objectives and the recognition of the impacts on the project results by unforeseen events or inadequate progress on specific Goals.
Simulation Tools & Techniques
Simulation can be used to tackle many problems, from initial design and proof-of-concept studies through operation and validation. Selecting the appropriate tool depends on the level of detail required for the model and the questions being answered. Some leading edge techniques include continuous simulation, program design, discrete-event simulation, combined discreet-continuous simulation and the Monte Carlo simulation technique. Less than a decade ago, simulation was developed using textual coding of the model. Today, graphical simulation tools are available. They are robust, reasonably inexpensive and allow for rapid model development with a reduced margin for error. They are also easier for non-technical staff to understand and require less training to learn to use than more complex tools.
Simulation refers to the broad collection of methods and applications to mimic the behavior of real systems. Today we have the ability to use the simulation techniques for creating a project roadmap and evaluating different options to achieve our end objectives. This activity done in the early stages of the project can help us create more realistic project plans and control processes. Exhibit -2 shows the process for creating models and using simulations for better forecasting.
In conclusion, all simulation analysis, for any situation shares the common elements of scenario comparison, risk identification and tradeoff evaluation. Some of the benefits of using simulation include: Comparison of Operating Alternatives by evaluating two or more alternatives, Time Compression by evaluate several different operating scenarios over a simulated period of time, Simulation and Optimization by imitating several situations with multiple levers and then optimize only a few that yield the best results, and Risk Management by developing mitigation strategies that will lessen the impact of unanticipated and disruptive events.
Some important things that need to be considered while using simulation techniques are:
- Level of detail (remember the forest v/s trees). More details mean more time, more bugs, and more parameters. It does not mean more accuracy
- Invalid Models: model vs. reality. Enough time should be spent in thinking about the kind of models that need to be created and would be useful in mimicking reality.
- Too short simulations. Create the simulations for the right intervals.
- Inadequate level of user participation. This is a common problem for all planning exercises!
- Obsolete or nonexistent documentation on requirements or past project plans.
The use of simulation techniques in projects increases the overall yields and quality of the Project to which they are applied. It also lowers the total program costs. Though it is a simplification of the real world, it is not a panacea as the predictive power of simulation is strongly dependent on how well the models are validated. Every Project has risk and uncertainty associated with it. Dynamic Project Management using simulation is a new trend that assures results by minimizing those risks and uncertainties through use of predictive models.
Knowledge of the different techniques and their application can help a project manager to use predictive analytics to forecast future probabilities with a far more acceptable level of reality v/s traditional methods.
If applied correctly it can help save many a project from disaster.
Choudhury, S. 2004. Project Management, New Delhi: Tata McGraw-Hill Publishing Company Limited
Cooper, R. G. (2006, April). Formula for Success Retrieved on July Stage-Gate Inc 27,.2006 http://www.stage-gate.com/
Global Project Design (2006) The Teamport™ Roadmap Design Process. Retrieved from http://www.gpdesign.com/teamport.html
Khwaja, A. (2005, September). Business Simulation Models – Weapons for Decision Making .Retrieved 07.10.06 from CGN & Associates. http://www.cgn.net.
Kleijnen, J. P.C. (2005). Supply chain simulation tools and techniques: a survey. International Journal of Simulation & Process Modeling, 1, pp82.
Moser B. & Milthemer, H.. (2003, February) Global Project Design LLC. White Paper - Planning in the Global Economy retrieved on July 20, 2006 from http://www.gpdesign.com. (GPD White Paper series 2003)
Mukhles Z. (2005, November). Predictive Analytics; the future of Business Intelligence Business Intelligence Research Center. http://www.vendorshowcase.com/Research/ResearchHighlights/BusinessIntelligence/2005/11/research_notes/TU_BI_XMZ_11_08_05_1.asp
Oren, T.I, Zeigler., B.P., Elzas, M.S. (1984) Simulation and Model-Based Methodologies: An Integrative View, NATO Scientific affairs Division, Tokyo. Published in cooperation with NATO Scientific Affairs Division
Piper, R. (2001). Putting Professionalism in to Project Management. Proceedings Project Management Institute - Europe. http://www.bis-co.com/pdf/PuttingProfessionalismIntoPM.pdf
PMI (2006b) Effective PMOs cross the enterprise. PMNetwork20(8) 12.
PMI (2006a) Doomed to fail. PMNetwork20(8) 16.
Product Development Institute Inc. (2001) Doing it Right-Winning with New Products Retrieved on July 18,.2006 from www.Synergest.com
Standish Group (2001, February/March) 1994 CHAOS Report, SofwareMag. Retrieved from http://www.softwaremag.com/archive/2001feb/CollaborativeMgt.html on July 12, 2006.
©2006, Animesh Arora
Originally Published as a part of PMI Global Congress Proceedings-Seattle, Washington