The Open Queuing Model software models Markov chains in applications of queuing theory. A Markov process is a statistical modeling tool named after Andrey Markov. It is essentially a system that can switch between many states, with the next state probabilistically depending only on the current state. OQM allows users to quickly build simulated queuing systems and predict the long-term behavior of that system. Based on the long term behavior the user can make adjustments to the system to optimize throughput, push traffic to certain areas, or balance the entire system.
A famous Markov chain is the so-called "drunkard's walk". Imagine a drunken person who is doing their best to walk along a line. They can either take one step forward or one step backward, with equal probability. This person is so drunk they don’t know where they are going and can’t remember where they were which makes the walk effectively random. From any point on the line, the drunken person has two possible transitions; forward a step, or back a step to the previous position. It turns out that many systems can modeled on this concept.
- Load balancing servers
- Predicting foot traffic flow at a venue
- Estimating page flow through a website
- Load testing a series of disks
- Predicting caller traffic through a call center
- Estimating document flow through a bureaucracy
- Model simple financial markets
- Build models in a simple spreedsheet format
- Save and reuse models with a simple .CSV format
- Open and modify models in any spreadsheet applicaiton
- Quickly run simulations and view node steady-states
- Break periodic systems with feedback loops
- Find the optimal feedback for convergence on any system
- Optimze systems by balancing work loads across nodes