Dynamic modeling

We develop dynamic computer models that help clients make better decisions, formulate robust strategies, and identify levers for performance improvement.

• Our approach provides insights into the key cause-and-effect relationships that drive business performance over time.

Applications

We have 15 years experience developing dynamic computer models for these business applications:

• Evaluating strategies under alternative future scenarios
• Solving complex problems that require understanding the dynamics of the business
• Identifying process improvements and testing them prior to implementation
• Analyzing investment decisions that involve uncertainty and risk

Approach

We use a collaborative approach to model development that focuses on the needs of the client. We concentrate initially on high-level models to quickly identify areas of opportunity and avoid unnecessary detail. Our goal is always to create the simplest, most elegant model possible that will provide the greatest return on your modeling investment. Other key elements of our approach include:

• Defining a clear purpose for the model and using it to maintain focus
• Utilizing the most appropriate modeling tool for the problem at hand
• Making all assumptions explicit and testable (avoiding black boxes)
• Employing visual mapping and animation to depict key relationships in the model
• Using a scientific approach to test and validate the model

Modeling toolkit

System dynamics is an efficient, elegant modeling technique that is ideally suited for creating high-level strategy models of complex businesses systems. The focus of system dynamics is on the interdependencies between system elements rather than the details within each element. System dynamics complements more detailed modeling techniques (e.g. discrete event simulations) by providing a high-level understanding of system behavior before investing time and resources in more detailed investigations.

We use discrete-event simulations to analyze and optimize business processes, including supply chains, operations, back-office services (e.g. IT project delivery), customer service, call centers, and healthcare processes. SIMUL8 software allows simplified process maps to be easily created, simulated, and animated to show process flows and queue levels over time. Discrete event simulations are particularly useful for identifying bottlenecks, resource constraints, quality problems and cost drivers that limit process performance.

In this approach to modeling, a complex system is represented by a collection of agents that are programmed with simple behavior rules. Agents can interact with each other and with their environment to produce complex collective behavior patterns. In contrast to system dynamics, which uses a top-down modeling approach with a high level of aggregation, agent-based modeling uses a bottom-up approach. We use agent-based modeling when it is simplest to render individual behavior in a model (e.g. customer, employee, product, patient, etc.) and the goal is to discover the collective behavior patterns that emerge. Applications include analyzing product development pipelines, employee development processes, social networks, pedestrian and traffic patterns, and marketplace dynamics.

Complex, real-world decisions about pricing, capital investments, R&D projects, etc. must be made without perfect information. To properly evaluate such decisions, and thereby avoid unnecessary risk, requires taking uncertainty into account. Maximizing the value created by investment opportunities also requires exploring possible intermediate decision points along the way. Decision analysis techniques include Monte Carlo simulations, decision trees and real option analysis. We can help you apply these techniques to analyze your most important decisions and explore a variety of strategies for minimizing risk and maximizing shareholder value.

Real world problems do not always match up well with a single modeling paradigm. Often, different aspects of a problem favor different methodologies. AnyLogic software combines the modeling approaches described above (system dynamics, discrete event, agent-based modeling) inside a single object-oriented platform. This provides us with the flexibility to apply the most appropriate combination of approaches for a given problem within the same model. For example, a multi-approach model could be used to explore the interactions between customers and employees (agents) within a customer service process (discrete event model) subject to a changing business environment (system dynamics model).