Internet network design specialists are looking for technologies and strategies to deliver network service under increased demand conditions. The choice of strategies is based on applying optimization and decision-making methods to select the most appropriate cable network design considering criteria established by the problem definition. However, this definition is itself a decision problem that has not received analysis in the literature. In particular, one of the most important questions is the necessity to define an expansion strategy. The first alternative (expansion) is to design a network to serve consumers with Internet demand equal to or greater than the predefined one to expand the network annually to serve consumers that reach the predefined Internet speed. The second alternative (oversizing) is to design a network to serve consumers with future Internet demand (after 5 years) at or above that the predefined one. Considering this, the objective of this research is to define the most advantageous strategy of expansion planning to attend a 5 years forecasted Internet demand, considering: (1) the possibility of utilizing a Gigabit-capable Passive Optical Network technology; (2) the application of the minimal Steiner tree and Dijkstra algorithms in planning procedures; (3) the influence of economic and technological factors on the demand forecast; (4) the aggressive, moderate, and conservative scenarios in decision-making. The results show that the over-dimensioning strategy reduces network investment by between 30 and 41%, but that this reduction does not always lead to a market investment ratio higher than that observed in the expansion strategy.

中文翻译：

---

扩大还是超大？需求增长场景下的互联网接入网络规划

Internet 网络设计专家正在寻找在需求增加的条件下提供网络服务的技术和策略。策略的选择基于应用优化和决策方法来选择最合适的有线网络设计，并考虑问题定义所建立的标准。然而，这个定义本身就是一个决策问题，在文献中没有得到分析。特别是，最重要的问题之一是定义扩张战略的必要性。第一种选择（扩展）是设计一个网络来为互联网需求等于或大于预定需求的消费者提供服务，每年扩展网络以服务达到预定互联网速度的消费者。第二种选择（规模过大）是设计一个网络，为未来互联网需求（5 年后）的消费者提供服务，达到或超过预定义的需求。考虑到这一点，本研究的目的是确定最有利的扩展计划策略，以适应 5 年预测的互联网需求，考虑：（1）利用具有千兆位能力的无源光网络技术的可能性；(2) 最小Steiner树和Dijkstra算法在规划程序中的应用；(3)经济技术因素对需求预测的影响；(4) 决策中的激进、温和和保守情景。结果表明，超维策略减少了30%到41%的网络投资，