Inventory control
After discussing the optimum re-order level and replenishment quantity calculations for single product situations, I have taken the possible interaction effects of a multiple products scenario into account. I have considered two separate scenarios, first the optimisation of economic order sizes when the limitation was set on the capital invested in stocks. I have introduced the Langrange multiplier to the problem and explored the significance of it. In the second scenario, I placed the emphasis on optimising economic order sizes when the limitation was set on total storage space.
Operational demand analysis
The operational demand analysis started with the development of an automatic data cleaning procedure for fault detection and diagnostic purposes. The idea of the development of such a program was to avoid tedious and time-consuming manual de-spiking. The important features of the program were not described in the thesis because I believe that it is outside the scope of this dissertation. Nevertheless, its application was successful.
In the daily demand and production analysis, I have defined the underlying trend of daily demand data and any in-flow pattern and related them to one another. I have identified and verified an appropriate auto-regressive 2 model for the averaged demand by using the autocorrelation and partial autocorrelation functions and indicated the goodness of fit of the second order auto-regressive model to the total daily in-flow and total daily demand time series.
I have constructed a linear model of road demand, fitted the model to the time series data and verified the adequacy of the model through the residuals.
Factors of the logistics strategy and system requirements
I have investigated the various logistics and related issues of a location choice for the establishment of new manufacturing facilities. I have quantified the sensitivity of a manufacturing location in terms of the associated logistics costs as a function of net added costs. Further, I have developed a complete corporate logistics audit system to allow organisations to assess their materials management systems' compliance to quality standards and advanced system requirements. My goal here was to foster the development of fundamental logistics systems that provide for continuous improvement, exceptional delivery performance and customer satisfaction.
5.2 Applicability of the results
Effective logistics performance measurement and control-ship are necessary to allocate and monitor resources.
As logistics competence becomes a more critical factor in creating and maintaining competitive advantage, accurate logistics measurement and control-ship increase in importance.
The logistical system performance analysis tools developed in the thesis focus on selected, main operational areas and provide concise analysis of these functions. First the emphasis was placed on the first element of the supply chain, purchasing. This section demonstrated the potential of using principal components analysis to reduce the analytical complexity of purchasing benchmarking studies. The original number of benchmark variables, 18, which represents a typical outcome of gathering key purchasing performance indicators, was finally reduced to 5, creating the opportunity for a simplified performance analysis. The result fulfils the industrial requirements of having the possible minimum number of factors to deal with, and yet expressing the most of the purchasing activity.
The second section provides an integer programming approach with branch and bound to the formulation of job-shop schedules, which are optimal in terms of the mean schedule flow-time. The applicability of inventory control analysis lies in providing different optimum calculations with business restrictions and re-order levels and replenishment order quantities when demand per unit time is variable but the lead time is fixed and when demand and lead time duration are both variable.
In the third chapter, not only were the “physics” of the demand investigated, but demand modelling techniques were studied. It is hoped that the results of the analysis will be useful in other applications and that the approach itself will have great applicability.
While no claim to optimality was made for the work on demand modelling, it is important to note that the optimal approach is always relative and application dependent.
Finally, chapter four developed a concise materials management system assessment at the corporate logistics level. The intention was to assess a materials management system’s conformance to the most advanced materials management system requirements.
5.3 Suggestions and future work
During this work a number of topics have been covered and many issues and questions raised. Some of these have been answered, others remain as challenges for the future.
The application of Markovian decision models to the job shop scheduling problem by introducing different maintenance states of machines and formulating a stochastic process that is a Markov chain with a known transition matrix, will be left for future work.
There are some issues that have not been addressed in chapter 3, when analysing and modelling operational demands. Specifically, the topics of applications of the Kalman filter to the time series measurement data, alternative statistical techniques, such as the sensitivity analysis, factor analysis and data extraction have not been considered. It has been assumed that the applied techniques are the most effective ones for the purpose of the thesis.
In chapter 4, due to lack of information from the examined companies (data confidentiality in most cases), a relative sensitivity analysis could not be conducted.
Having a nominal value for indices such as annual value of products produced, value of raw materials purchased, raw material transport costs, raw material handling costs, and transport of manufactured products would make it possible to evaluate normalised measurements of optimality.
However, the challenges will never stop due to the evolving nature of logistical operations. There is hardly any activity that could be improved or take on a new perspective from better analysis of its operations.
As the continuation of my work, the plan is two fold. Firstly to provide numerical examples for those methods examined in the thesis that have not been underlined by a case study. Secondly, to prepare an operational logistics benchmarking study.
Operational benchmarking in a sense that it focuses on the activities and processes that perform the basic business functions. More specifically, a function-wide benchmarking has been planned, which will involve simultaneously assessing all the tasks in a logistics function. The first identified logistics function is warehousing with the aim to improve performance including operations from storage, put-away, order selection, re-warehousing, picking and shipping.