Study On A New Last Time Buy Quantity Forecast Model Based On Failure Rate And Active Service Unit In The Company A

As competition in PC industry becomes more and more seriously, margin in thisfield becomes thinner and thinner. To survive, a company should cut down cost asmuch as possible while keeping high serve level. For achieving this, it should have agood demand forecast first. Last time buy spare parts forecast, as a part of demandforecast, is one of most difficult and important sections. Currently, among all forecasttechnologies which can be applied in last time buy forecast, there is no a method thatcan meet high robustness, high accuracy and high practicality these three requirements.Therefore, PC industry urgently need a new, three metrics fulfilled last time buyforecast model.In the paper, it firstly analyzed pros and cons of all forecast methods which couldbe used in last time buy forecast around world, and found that a failure rate and activeservice unit based method was a good candidate. Secondly, it gave effective solutionsfor disadvantages in the traditional last time buy forecast model, for example, lowaccuracy failure rate data and active service unit data processing methods, and lowaccuracy failure rate and active service unit multiplication method, etc. In detail, byemploying past six months actual failure rate data and Weibull distribution method, itgot more accurate failure rate forecast; by introducing active service unit matrix, itgained more specific active service unit; by multiplying more accurate failure rate andmore specific active service unit with new methodology, it enjoyed VBA programmingadvantages; by using non-linear programing relevant theory, it successfully establishedan improved last time buy forecast model. Finally, it gave a case study in the targetcompany, validated the effectiveness of the new model and analyzed the process ofsucceeding to implement the new model which built foundations for other companiesin PC industry to use it.