The Energy-Delay Pareto Front in Cache-enabled Integrated Access and Backhaul mmWave HetNets (2201.08308v1)

Abstract: In this paper, to address backhaul capacity bottleneck and concurrently optimize energy consumption and delay, we formulate a novel weighted-sum multi-objective optimization problem where popular content caching placement and integrated access and backhaul (IAB) millimeter (mmWave) bandwidth partitioning are optimized jointly to provide Pareto efficient optimal non-dominating solutions. In such integrated networks analysis of what-if scenarios to understand trade-offs in decision space, without losing sight of optimality, is important. A wide set of numerical investigations reveal that compared with the nominal single objective optimization schemes such as optimizing only the delay or the energy consumption the proposed optimization framework allows for a reduction of the aggregation of energy consumption and delay by an average of 30% to 55%.

• The paper presents a multi-objective optimization framework that jointly optimizes caching placement and mmWave integrated access/backhaul to achieve a Pareto efficient trade-off between energy consumption and delay.
• Numerical results indicate that caching time, caching capacity, and small cell density critically influence performance, with increased caching time reducing delay while raising energy consumption.
• The study demonstrates that adjusting the weighting parameter enables network operators to choose optimal operating points, significantly outperforming single-objective optimization schemes.

Pareto Optimization for Energy and Delay in Cache-Enabled mmWave HetNets

This paper addresses the critical problem of backhaul capacity limitations in heterogeneous networks (HetNets) by introducing a multi-objective optimization framework that jointly optimizes content caching placement and integrated access and backhaul (IAB) millimeter wave (mmWave) bandwidth partitioning. The goal is to minimize both energy consumption and delay, providing a Pareto efficient set of non-dominated solutions that offer valuable trade-offs for network operators.

System Model and Problem Formulation

The paper considers a two-tier HetNet topology (Figure 1) consisting of a macro base station (MBS) with fiber backhaul and multiple small cell base stations (SBSs) with limited cache storage. Users request content from a popular content set, following a Zipf-like distribution. The key innovation lies in the integration of IAB mmWave technology, where the same spectrum is used for both access and backhaul links. The spectrum is divided into two portions (Figure 2): one for access links shared by all BSs and the other for orthogonal backhaul links between the MBS and SBSs.

Figure 1: The IAB millimeter (mmWave) heterogeneous networks network topology considered in this work.

The paper formulates a weighted-sum multi-objective optimization problem to minimize energy consumption and delay (Equation 23). The weighting parameter $\alpha$ allows for adjusting the relative importance of each objective, and normalization factors $\delta_e$ and $\delta_d$ are introduced to account for differences in magnitude between the objectives. The optimization problem includes constraints on caching decisions, cache capacity, bandwidth allocation, and quality-of-service (QoS) requirements. The problem is solved using MOSEK solver and CVX with MATLAB.

Numerical Results and Analysis

The numerical investigations explore the trade-off between energy consumption and delay under various network configurations.

Impact of Caching Time

The caching time, which represents the files' life-cycle, significantly affects the system performance. The results (Figure 3) show that the energy consumption increases while the delay decreases as the caching time increases, highlighting the competitive nature of these objectives. Comparing the proposed scheme with delay-only and energy-only optimization schemes, the results show that the sum of energy consumption and delay at the extreme points are always higher than the rest of the points of the same Pareto curve.

Figure 3: Caching decision.

Effect of Caching Capacity

The caching capacity of SBSs also plays a crucial role. As shown in (Figure 4), the amount of files cached at SBSs and the bandwidth portion allocated to the access link decrease with the increased energy consumption preference.

Figure 4: Caching decision.

Influence of Small Cell Density and Transmission Power

The small cell density and transmission power are critical parameters influencing network performance. As shown in (Figure 5), the overall performance improves with increased density until a saturation point is reached, where increased interference leads to higher energy consumption and network degradation. The scheme without edge caching capability generates the worst performance due to the longest delays in fetching content to the end users.

Figure 5: Aggregated delay.

Conclusions and Implications

This work provides a comprehensive analysis of the energy-delay trade-off in cache-enabled IAB mmWave HetNets. The multi-objective optimization framework allows for dynamic bandwidth partitioning and edge caching allocation, enabling network operators to select optimal operating points based on their specific requirements. The numerical results demonstrate that simply upgrading network capacity, the density of small cells, or the maximum transmission power may not always improve network performance. The actual life-cycle management of popular content needs to be considered when caching at the edge. The proposed framework achieves significant performance improvements compared to single-objective optimization schemes by appropriately selecting the weighting parameter $\alpha$.