3.4 Simulation Results and Discussion
3.4.2 Results and Discussion
The proposed JPRA algorithm is compared with the SSDT [14] and the LPPA schemes with the best-effort rate allocation. The major idea of the conventional SSDT algorithm is to dynamically choose one base station with the best link quality in the active set for transmission in order to mitigate interference caused by multiple-site transmission. However, because of the maximum link power constraint, SSDT may not be able to offer enough required power for soft handoff users with high transmission rates. Also, under the constraint of maximum link power, best-effort rate allocation is to assign handoff users maximum allowable transmission rate which exists feasible solutions of power allocation to satisfy users’
required quality of services. In order to achieve fair comparison for all schemes, the total power constraint of soft handoffs is confined to 0.3 times maximum transmission power of each base station. In the following, we take best-effort rate allocation as the benchmark for comparisons, and denote SSDT with best-effort rate allocation and LPPA with best-effort rate allocation by SSDT and LPPA, respectively
Figure 3.4 shows average handoff forced termination probability under different traffic load situations. It can be seen that, when the traffic load is light, the LPPA scheme improves
over the SSDT scheme. However, when there are more data users served in the system, the handoff forced termination probability of the LPPA scheme is worse than that of the SSDT scheme. The reason is that higher interferences are induced by the multi-site transmission mechanism than by the single-site transmission mechanism for handoffs. The results mean that power balance and power saving are important characteristics for the radio resource management of the handoff mechanisms, and both are impact factors on the system per-formance under different traffic load situations. Furthermore, it is found that the handoff forced termination probability of JPRA is superior to that of the SSDT schemes by over 300%. Besides, JPRA improves LPPA by around 200%, which can be inferred that the gain comes from the rate allocation by evolutionary computation method. The allowable transmission rates are highly coupled to the feasible solutions of power allocation, which can be formulated as a constraint optimization problem. The ECRA method is thus designed to find an optimal rate allocation for multirate soft handoff users, in which the goal is to accomplish maximum throughput of all soft handoff users such that the total power con-straints of soft handoffs for all base stations should be satisfied. Besides, because the rate allocation of each soft handoff user directly affects the management of power resource in at least two base stations, an optimal rate allocation for soft handoff users can further enhance the effect of power balance among cells.
Moreover, consider the case with measurement errors during the active set selection, it is observed that measurement errors incur higher handoff forced termination probabilities for all schemes because base stations waste more power on multirate handoff users. The improvement of JPRA is by around 500% over SSDT. Besides, because of link power con-straint and the single-site transmission mechanism, SSDT is more sensitive than JPRA and LPPA to the occurrences of measurement errors. In this case, it is found that the effects of measurement errors can be relieved by the multi-site transmission mechanisms and their power balance characteristic. With the superb power balance characteristic, not only can JPRA provides better service continuity performance but also possess the capability of the resistance to measurement errors. It is particulary noteworthy that the performance of the
3 4 5 6 7 8 9 10 11 12
Number of data users per cell
Total handoff throughput (kbits)
Figure 3.5: Total handoff throughput versus the number of data users per cell without measurement errors (ME) and with 1.5 dB measurement errors (ME).
handoff forced termination probability can also be regraded as the performance index of the cell’s service coverage, in which smaller handoff forced termination probability means better cell coverage. Thus, Fig. 3.4 also shows that JPRA achieves better cell’s service coverage than SSDT and LPPA.
Based on the viewpoint of the capacity, Fig. 3.5 shows the results of the total handoff throughput versus different number of data users. It is found that both LPPA and JPRA have higher throughput than SSDT, because of the multi-site transmission mechanism. Besides, due to total power constraint of soft handoffs, there exists the tradeoff between coverage and capacity for the LPPA and JPRA schemes. From Fig. 3.4 and Fig. 3.5, the results show that the ECRA method in the JPRA reduces average transmission rate of the multirate soft handoffs so as to accomplish better cell coverage while LPPA leads to more terminated handoff users but achieves higher handoff throughput than JPRA. The similar results could be observed for the case of measurement errors. Since more power is wasted by measurement errors, higher handoff forced termination probability makes more power left for the survived handoff users to transmit with higher average transmission rates. In addition, we can see that the ECRA method in JPRA plays an important role to reduce performance degradation
3 4 5 6 7 8 9 10 11 12
Number of data users per cell
Average call dropping probability
Figure 3.6: The average call dropping probability versus the number of data users per cell without measurement errors (ME) and with 1.5 dB measurement errors (ME).
by measurement errors. This is because the ECRA method allocates optimal transmission rates for multirate soft handoffs to further balance power loads among cells. Therefore, JPRA successfully enhances cell’s service coverage and handoff throughput.
In the following, we show that joint power and rate allocation strategies for multirate soft handoff have significant impacts on system performance in terms of average total throughput and average call dropping probability. Fig. 3.6 shows that JPRA can improve the average call dropping probability of all users by 71% and 200% without and with measurement errors compared to the SSDT scheme, in which the call dropping occurs when there is no feasible solution of power allocation to support users with their required service qualities for a period of time. From the preceding results of the handoff performance, the superiority of JPRA over SSDT is mainly because JPRA owns the power balance characteristic resulting from LPPA and ECRA methods. Since proper power balance can prevent one base station from wasting too much power resources to serve multirate soft handoffs, the power resource can be preserved to serve non-handoff users with higher transmission rates by optimally managing radio resource of multirate soft handoffs.
Fig. 3.7 shows the system performance of the total throughput gain referred to the SSDT
3 4 5 6 7 8 9 10 11 12
Number of data users per cell
Total throughput gain (refer to SSDT)
SSDT w/o ME
Figure 3.7: The total throughput gain, which is referred to SSDT, versus the number of data users per cell without measurement errors (ME) and with 1.5 dB measurement errors (ME).
scheme. We can see that JPRA can enhance the average total throughput than SSDT by 5%
and 8% in the case of measurement error free and 1.5 dB measurement errors, respectively.
From above results of system performance, it is found that JPRA makes great improvements of cell’s service coverage and system capacity because it can optimally allocate radio resource of soft handoffs.
In the meantime, on the perspective of the user satisfaction, voice and data users should have different service requirements. Denote the call dropping probabilities of voice and data users as Pv and Pd, respectively. Also, the summation of the allocated and required transmission rates of all data users are represented as Rdand R∗d, respectively. We then define two satisfaction indexes for voice and data users, denoted by USIv and USId, respectively, as
½ USIv = (P∗v− Pv)/P∗v,
USId = κd× Rd/R∗d+ (1 − κd) × (P∗d− Pd)/P∗d, (3.14) where κd and 1 − κd are the weighting factors for the total throughput and call dropping probability of data users, respectively. Here, both P∗v and P∗dare set as 0.05. For voice users, because of constant transmission rate, the satisfaction comes from call dropping probability.
3 4 6 8 10 12
Number of data users per cell (a) USIv
Voice user satisfication index ( USI v )
3 4 6 8 10 12
Number of data users per cell (b) USId
Data user satisfication index ( USI d ) SSDT w/o ME
Figure 3.8: The user satisfaction index (USI) versus the number of data users per cell for (a) USI of voice users (USIv) and (b) USI of data users (USId), respectively.
Also, for data users, assume κd is 0.7 because data users usually are satisfied with higher transmission rate and can tolerate longer transmission delay because of call dropping events and retransmission mechanisms. In Fig. 3.8, it is shown that the proposed JPRA scheme can provide outstanding user satisfaction indexes for voice and data users even when there exists measurement errors during active set selection.
For the mixed-size cellular system with mix-sized cells, microcells are normally congested cells and with stringent power budget of link power and maximum transmission power. It is useful to adopt multi-site transmission mechanisms for soft handoffs to balance some power loads into neighboring cells. Besides, because the constraint of maximum link power is tight for microcells, the single-site handoff mechanism, like SSDT, may fail to support soft handoff user enough required power for high rate services by the best link to microcells. Therefore, a reasonable inference is that when the cell radius ratio between macrocell and microcell gets smaller (ρ < 1/2), the larger gain could be obtained from the JPRA algorithm because of the outstanding power balance characteristic.