簡化操作型區間數線性規劃在水質管理之應用

一般線性規劃方法中，限制式之係數及右端常數皆爲確定數值，因此最佳解也爲確定數值。在真實的世界中，許多問題皆具有不確定性，而無法以確定數來表示，環境系統亦不例外。因此，有些研究利用序率線性規劃模式來探討此問題，然而序率方法需要有足夠的資料去檢定資料之機率分佈，故在資料不足時較不可行。本文簡化具有穩定性的操作型區間數線性規劃，應用在頭前溪河川水質的優化控制上。傳統的河川水質優化管理問題，其考慮的污染排放量以及河川流量皆爲定值，而求解出的最佳容許排放量（河川涵容能力）也爲一定値，這並不符合實際情況，因爲在真實世界中存在具有不確定性與難以確定數量化的元素。因此本文以區間數的方式表示不確定參數，則求出的結果亦爲一區間數，也就是一個容許的排放範圍，在管理上可提供更多的資訊，以訂定排放標準。

關鍵字

不確定性；涵容能力；優化

並列摘要

In the traditional algorithm of linear programming, coefficients and right hand side are constant. Thus, optimal solutions are also constant. However, there is uncertain information in the real world. It is not proper to express uncertain information as deterministic value. Stochastic linear programming is often used in solving problems with uncertainty, but it needs more data to identify the probability distribution of concerned information. When data is limited, it is infeasible. In this study, simplified operational interval number linear programming (SOINLP) is used to solve an optimization problem of water quality control for the Touchien Creek. The traditional linear programming model for water quality management has constant streamflows and pollution discharges, which produces constant permissible pollution discharge (assimilation capacity). It is not realistic, because there is uncertainty in the real world. Thus, interval parameters are used in this study, and optimal solutions with interval values of decision variables are produced by the SOINLP. It offers more information for water quality management.

並列關鍵字

Uncertainty ； Assimilation capacity ； Optimization

主題瀏覽