Forecasting of compound Erlang demand

摘要: Intermittent demand items dominate service and repair inventories in many industries they are known to be the source of dramatic inefficiencies defence sector. However, research forecasting such has been limited. Previous work this area developed upon assumption a Bernoulli or Poisson arrival process. Nevertheless, intermittent patterns may often deviate from memory-less assumption. In we extend analytically previous important results model based on compound Erlang process, provide comprehensive categorisation scheme used for purposes. numerical investigation assess benefit departing insights into how degree determinism inherent process affects forecast accuracy. Operationalised suggestions offered managers software manufacturers dealing with items.