This paper chooses a Data Envelopment Analysis (DEA) model for different areas to identify the difference in solving the industrial pollution problem by comparing their levels of efficiency. Generally, the industrial pollution problem calls for within-area treatments, although these affect areas beyond their limits. Following this fact, a Maximal Efficiency Sum (MES) DEA Model is used to estimate the anti-industrial pollution efficiency of different cities in Anhui province of China.

A model of scenario-based line capacity expansion problem for PWB (Printed Wiring Board) assembly systems at the aggregate level is developed. The model synthesizes BOM (Bill Of Material) of product families and machine operation flexibility, thus it is an attempt of integrating strategic capacity planning, aggregate production planning and MPS (Master Production Scheduling), which is an important research topic of production management. Since the resulting model is a large-scale two-stage stochastic mixed integer programming problem, it can not be solved with standard code. An approximate solution procedure is developed, which first reduces the searching space of capacity expansion decision variables to rough addition sets by heuristics, then the rough addition sets are searched through adaptive genetic algorithms. Numerical experiments are presented to show the financial benefit of the model and the feasibility of our approach.

In a multi-product, flexible manufacturing environment, line capacity of printed wiring board (PWB) assembly systems may need to be adjusted at the beginning of each aggregate planning period because of demand fluctuation over multiple periods. A model of production planning and equipment changeover scheduling at the aggregate level is developed. In the described model, three kinds of equipment changeover methods, i.e. adding machine, removing machine and transferring machine, are involved. Because the model is a large-scale integer programming problem, it cannot be solved directly. A solution approach is developed, which first solves a recursive linear programming problem to obtain a rough set of machines to be added and a rough set of machines to be removed for each machine line in each period, then applies a branch and bound heuristic to the rough sets to obtain near-optimal solutions to the equipment changeover scheduling problem. Computational studies show the financial benefit both on capital cost and equipment changeover costs.

As a result of the fierceness of business competition, companies, to remain competitive, have to charm their customers by anticipating their needs and being able to rapidly develop exciting new products for them. To overcome this challenge, technology forecasting is considered as a pow-erful tool in today's business environment, while there are as many success stories as there are fail-ures, a good application of this method will give a good result. A methodology of integration of patterns or lines of technology evolution in TRIZ parlance is presented, which is also known as TRIZ technology forecasting, as input to the QFD process to design a new product. For this purpose, TRIZ technology forecasting, one of the TRIZ major tools, is discussed and some benefits compared to the traditional forecasting techniques are highlighted. Then a methodology to integrate TRIZ technology forecasting and QFD process is highlighted.

In a multi-product, flexible manufacturing environment, line capacity of printed wiring board (PWB) assembly systems may need to be designed at the beginning of each aggregate planning period because of demand #uctuation over multiple periods. A model of line capacity design problem and production planning at the aggregate level is developed, in which production and subcontracting are assumed to be two options for a firm to meet market demand. The model presented is a large-scale integer programming roblem, it cannot be solved by using standard-or mixed-integer programming codes. Under the ssumption that each machine line is dedicated to produce one product family, the model can be decomposed as a relatively small subproblem, and each subproblem has good properties by which the subproblems can be further simplified and decomposed over multiple planning periods. As the result, the original large-scale two-stage integer programming problem can be approximately solved by solving a series of small-scale mixed-integer programming, which can be implemented on a workstation or a PC. Computational studies show that the solution method is developed which gives near-optimal solutions with much less computational errort.