Joint resource allocation and base-station assignment for the downlink in CDMA networks

In this paper, we jointly consider the resource allocation and base-station assignment problems for the downlink in CDMA networks that could carry heterogeneous data services. We first study a joint power and rate allocation problem that attempts to maximize the expected throughput of the system. This problem is inherently difficult because it is in fact a nonconvex optimization problem. To solve this problem, we develop a distributed algorithm based on dynamic pricing. This algorithm provides a power and rate allocation that is asymptotically optimal in the number of mobiles. We also study the effect of various factors on the development of efficient resource allocation strategies. Finally, using the outcome of the power and rate allocation algorithm, we develop a pricing-based base-station assignment algorithm that results in an overall joint resource allocation and base-station assignment. In this algorithm, a base-station is assigned to each mobile taking into account the congestion level of the base-station as well as the transmission environment of the mobile.     

Non-newtonian squeeze films in journal bearings

Non-newtonian behaviour, which occurs in various lubrication problems, was studied using a power law model. The geometry considered was that of a journal bearing. Bearing characteristics i.e. the load capacity and the response time, were obtained for full-journal and half-journal bearings and were compared with each other. As the flow behaviour index n increases the load capacity ratio and the response time ratio decrease.     

Hydrodynamic performance of steelmaking tundish systems: a comparative study of three different tundish designs

Extensive water modelling was carried out to ascertain the influence of various types of baffle designs on the hydrodynamic performance of three different designs of steelmaking tundish systems. These included, a two-strand slab casting tundish, a six-strand billet casting tundish and a five-strand, skewed, delta shaped tundish. Plant scale operating conditions were scaled down respecting both geometric and Froude similarity and on the basis of the latter, the inflow rate of water into the model tundish systems was estimated via: Q_m = Λ^5/2Q_f.s. To quantify the hydrodynamic performance, residence time distribution (RTD) characteristics were measured using the conductivity measurement technique for a wide range of baffle designs. From such measurements as well as from flow visualisation studies, the following general observations have been made. The optimum design of baffles together with its number and position within the tundish appear to be a strong function of the basic tundish design (viz., the geometry, the number of operating strands etc.). Of the various types of baffles investigated in this work (dam, dam + weir, baffles with holes etc.), appropriately designed slotted baffles appear to modify the RTD characteristics most favourably towards superior metallurgical performances. Increase in the number of strands, asymmetricity in the tundish design and flawed operating conditions, (viz., large width to length ratio) were all found to influence the tundish performance adversely. A comparison of experimental results for the three tundish systems indicated that changing the characteristics of the baffle design can lead to significant performance improvement in the case of the two symmetric tundish systems (e.g., the two and the six strand tundishes), the five strand skewed, delta shaped tundish was, however, found to be somewhat insensitive to such changes. For such a tundish geometry, no design of the flow-modification devices tried, could bring the hydrodynamic performance any closer to the best results obtained for the two other tundish systems. Possible reasons for such observations are discussed in the text.     

Determination of Test Intervals in Certain Repairable Standby Protective Systems

A System has associated with it a Safeguard such that whenever a failure in the Safeguard is detected, the System is shut down. The System is restarted as soon as the Safeguard is repaired. An example is a standby protective system attached to a nuclear power plant; as soon as failure is detected in the protective system the nuclear plant will be shut down until repair is completed. In order to insure that the Safeguard remains in the utmost state of readiness, the components of the Safeguard are periodically tested, and repaired if the test reveals a failure. In this paper we define a measure of readiness appropriate for the above systems and consider the problem of determining the testing frequency of the components of the Safeguard in such a way that an adequate level of readiness is maintained. As an illustration, the method is applied to a typical reactor-trip system.     

Objective-driven monitoring for broadband networks

An approach to sensor configuration, installation, and activation for real-time monitoring of broadband networks for performance management is presented. An objective-driven measurement strategy for establishing the dynamic and statistical databases of the network is described. Objective driven monitoring allows the activation of sensors for data collection and abstraction based on a set of objectives. The objectives are derived from the quality of service requirements for real-time traffic control and operator submitted queries. The methodology of objective-driven monitoring for selective activation of sensors is implemented as a set of rules in the knowledge base of the monitor     

Partial removal of hafnium from zirconium through the reactions of tetrachloride vapours with solid sodium chloride

A pyrometallurgical method for partial separation of hafnium from zirconium has been investigated. The method involves continuous reaction between a flowing gas mixture of zirconium tetrachloride and hafnium tetrachloride with solid sodium chloride in a packed bed reactor. The solid-vapour reaction enriches the vapour phase with respect to zirconium tetrachloride. This may be attributed to a more favourable reaction for hafnium tetrachloride. The influence of a number of parameters, e.g. temperature, gas flow rate, particle size and bed length, has been studied. Decreases in flow rate and particle size and increase in bed length are found to favour separation, while temperature has little effect. A simple equation has been derived to explain the experimental observations.     

Some Bayes Estimates of Long-Run Availability in a Two-State System

Some Bayes estimates are obtained of an index of performance of a system that alternates between two states, up or down, in accordance with a Markov process. The index considered is long-run availability, which measures the probability that the system will be up when needed. For the purpose of obtaining these estimates, two types of observations are considered: those that reveal only the state of system at isolated time points and those that continuously record the duration of the up and down times of the system.     

Randomized trial of etoposide and cisplatin versus etoposide and carboplatin in patients with good-risk germ cell tumors: a multiinstitutional study

PURPOSE: This multicenter, randomized phase III clinical trial evaluated the efficacy of etoposide plus carboplatin (EC) versus etoposide plus cisplatin (EP) in good-risk germ cell tumor (GCT) patients. PATIENTS AND METHODS: Between October 1986 and December 1990, 270 patients with good-risk GCTs were randomized to receive four cycles of either EP or EC. The etoposide dose in all patients was 100 mg/m2 on days 1 through 5. EP patients received cisplatin at 20 mg/m2 on days 1 through 5 and therapy was recycled at 21-day intervals. For EC patients, the carboplatin dose was 500 mg/m2 on day 1 of each cycle and the EC recycling interval was 28 days. RESULTS: Two hundred sixty-five patients were assessable: 131 patients treated with EC and 134 treated with EP. One hundred fifteen of 131 assessable patients (88%) treated with EC achieved a complete response (CR) versus 121 of 134 patients (90%) treated with EP (P = .32). Sixteen patients (12%) treated with EC relapsed from CR versus four patients (3%) treated with EP. Therefore, 32 patients (24%) who received carboplatin experienced an event (incomplete response [IR] or relapse) compared with 17 of 134 patients (13%) who received cisplatin (P = .02). At a median follow-up of 22.4 months, event-free and relapse-free survival were inferior for patients treated with EC (P = .02 and P = .005, respectively). No difference in overall survival was evident (P = .52). CONCLUSION: Two-drug therapy with EC using this dose and schedule was inferior to therapy with EP. Cisplatin remains as the standard platinum analog in the treatment of patients with good-risk GCTs. Carboplatin should be restricted to investigational trials in GCT.