This PR changes the random time between consumer visits from a uniform distribution to a exponential distribution. This is a more realistic distribution for consumer visits.
The number of consumer per minute is fixed. Previously, with the setting of 100 consumers per minute, only ~32 arrived per minute. The calculated mean waiting time was to high. Also, the consumer's processing time is now considered when calculating the waiting time. After these changes ~100 consumers arrived per minute (sometimes more, sometimes fewer) in my tests.
The min_wait and max_wait option is no longer needed.
I set a random seed to have a comparable consumer behavior between simulations.
Added the Gemfile.lock with an up-to-date nokogiri version to fix the Github warning.
The unused library statsample-glm was removed.
This PR changes the random time between consumer visits from a uniform distribution to a exponential distribution. This is a more realistic distribution for consumer visits.
The number of consumer per minute is fixed. Previously, with the setting of 100 consumers per minute, only ~32 arrived per minute. The calculated mean waiting time was to high. Also, the consumer's processing time is now considered when calculating the waiting time. After these changes ~100 consumers arrived per minute (sometimes more, sometimes fewer) in my tests. The min_wait and max_wait option is no longer needed.
I set a random seed to have a comparable consumer behavior between simulations.
Added the Gemfile.lock with an up-to-date nokogiri version to fix the Github warning. The unused library statsample-glm was removed.