Accelerating process excellence using virtual discrete event process simulation |
About the project
Figure 1 [left] illustrates the key elements of the intended modeling environment. The cycle of events occurring will begin with V-DEPS model creation/updating. The discrete event simulation process will then be undertaken with the option of optimising specific elements of the model, eg minimise work-in-progress. Automatic analysis of the simulation results will then take place and will be used to identify a prioritized list of waste root causes and generate a set of performance metrics. Creative problem solving support will then be provided for solving specific problems. At all stages advanced visualization of the processes and performance metrics will be provided as well as lean training facilities. The technical approaches to the development of the overall system are as follows:
1. Development of a V-DEPS modeling tool capable of virtual modeling of complex processes, complex systems and the activities involved in lean practices at both process and system levels. At present traditional virtual process designs focus on modelling processing functionality whereas lean improvement characteristics within V-DEPS would need to include the ability to make changes for example to process set-up activities, a workplace’s equipment and materials layout arrangements, and the layout of the processing facilities within a work system. In addition, performance metrics would need to be provided by which initial and changed states could be compared. Hence, discrete event simulation functionality needs to be provided to a greater range of modelling elements and at higher levels of operational detail. The development of the V-DEPS modelling capability will be achieved by identifying a comprehensive list of lean practices through literature search and industrial surveys, undertaking a gap analysis, ie comparing this list with current capabilities of V-DEPS systems, and developing V-DEPS modeling functionality to fill the gaps identified.
2. Development of an Advanced Visualization capability which is responsive to user needs and enables users to physically interact with V-DEPS models and understand the complexity of their behaviours. In addition, visualistion technologies must be developed to allow the learner to assimilate quickly the necessary information during the progress of the simulation. Lean improvements are often made at the work place level whilst their benefits must be obtained at the systems and organisational levels. This visualisation function will need, therefore, to be provided at work place, work systems and organisational levels. Here discrete event simulation has the ability to store large amounts of data with respect to the complex dynamic operational behaviours of work systems before and after improvement changes have been made. In addition, at the end of each simulation run data concerning the operational efficiency of the work system is stored. Methods must be developed for displaying this data in a manner that enables more experienced lean practitioners to determine the true benefits of their changes/decisions in a simulated real time environment. Such simultaneous visualisation of the workplace, systems and organisational views are essential in work environments where sequential activities are carried out in separate departments and/or buildings such as in a hospital environment. Also of importance will be to ensure the presentation and visualisation of high-level management information. Standard advanced visualization hardware, based on the Virtalis StereoWorks system, will be employed with the project work focusing on identifying the information needs for presentation to users and developing methods by which this can take place.
3. External Data Input/Output (I/O) tool capable of inputting external data such as customer demand requirements and model modifications to V-DEPS and outputting details of required improvement activities.
4. Development of Self-Organising Relationship Matrices. In order to represent the complexity of relationships within a V-DEPS model it will be necessary to identify the various type s of interactions that take place between the various types of modeling elements. These interactions will include those between physical modeling elements, (eg processing equipment and inventory queues), those between the rules used to control the sequence of activities that take place, (eg priority rules for selecting jobs to progress), and those associated with cultural/change management aspects. These relationships will be represented, initially as networks, but then converted to Relationship Matrices in order that the Optimisation and Simulation Management tools can adapt these relationships such that external environmental changes are allowed to change the behaviour of the model. Each matrix will represent the network of relationships between two types of modeling elements. An individual relationship matrix will evolve from the interactions of individual modeling elements during a simulation run. Interrelationships between the main categories of modeling elements will include
- individual root causes and their resulting effects,
- root causes and problem solving team requirements,
- root causes and the lean enablers required to permanently solve them,
- V-DEPS modeling elements and root causes,
- lean enablers and lean/change management skills,
- lean/change management skills and training requirements,
- and lean/change management skills and improvement team organization.
Generic relationships within each of these categories will be identified through a combination of literature review and industry surveys. Methods will need to be developed to enable the actual relationships that exist within an individual model to be identified.
5. Development of an Optimization tool that will automatically identify optimum solutions from amongst the alternatives proposed during the creative problem solving process. The main technique to be employed will be genetic algorithms (GAs) which also possesses the facility to create new solutions to problems. The GA processes developed must be capable of adapting V-DEPS models to meet changing external environmental needs.
6. Simulation Management tool which is capable of planning and controlling individual simulation runs in terms of deciding alternative model designs and parameter value ranges to be examined and automatically generating and running these models. This functionality must be closely integrated with the Optimisation tool since the application of genetic algorithms creates alternative models the effectiveness of which must then be identified through simulation.
7. Performance and Results Analysis tool that is capable of analyzing the output from simulation runs and identifying root causes of productivity and resource wastes such as excessive buffer sizes and long cycle times. It is important that the simulation and results analysis run in background mode to enable responsive visualization of the results of model changes. Simulations generate large amounts of data, hence identifying emergent behaviours will represent a challenging task and will require the development of self-organising data mining processes.
8. Development of a Creative Problem Solving (CPS) tool capable of leading improvement teams through the CPS process from the initial selection of improvement team members based on ensuring that appropriate Continuous Improvement (CI) and Team Leadership skills are available within a team to the final selection of the best solution to adopt. This will require identification and presentation of information within the modelling environment that will enable learners to build-up their creative problem solving and team leadership skills. For more experienced lean practitioners building up their creative problem solving skills enables more innovative lean solutions to be identified. Here we need to understand what modelling functionality is required to support the creative problem solving process particularly when multi-skilled teams are required. Principles from industrial psychology underpinning team working, change management and motivation will play an important role. The basic processes involved in lean problem solving are well known. Hence, this task will involve automating these processes and making them available along with appropriate data and simulation results through the visualisation environment such that team-based autonomous decision making can be enabled.
9. Development of a Lean Training environment aimed at improving both the CI skills of lean practitioners and the process and operations skills of those required to implement and operate the lean improvements. Standard training scenarios will be developed for the main lean enabling methods. These scenarios will provide visual cues to lead new learners through individual lean change processes. Learners will be expected to undertake, using the V-DEPS environment, data collection tasks, data analysis tasks, generation of alternative solutions, selection of best solution and solution implementation tasks. Since the V-DEPS tools will be able to model an organisation’s actual work environment then lean training environments can be specifically developed for these. Through use of the ‘lean/change management skills and training requirements’ relationship matrix the automatic identification of skills gaps will be possible and hence the self-organisation of training requirements.
10. Change Management tool aimed at enabling the human resource issues associated with lean practices to be included in the formation of teams, design of lean solutions, analysis of alternative lean solutions and the selection of the most appropriate solution to implement. For example, this would take into account relationships between team skills, team and organizational roles and responsibilities, motivation and team performance.