Search Documents

Roof falls in coal mines typically occur in or around intersections and this has not changed in the last two decades. This research develops an improved scientific understanding of stress distribution and associated failure behavior around 4-way and 3-way coal mine intersections. Three-dimensional numerical analyses were performed to determine factors that influence intersection stability. The analyses used Hoek-Brown failure criterion. Failure zones around 3-way and 4-way intersections were developed. Intersection span and horizontal stress have a major influence on intersection stability. For the 4-way intersection, pillar corners across the intersection fail first and lead to progressive failure of immediate roof and floor layers. The mechanism of failure is similar for the 3-way entry but the shape and extension of failed zones differ. Coal ribs mostly fail due to tensile stress, while roof and floor strata fail due to shear stresses. Rib corners fail due to a combination of shear and tensile stresses. In addition to stress distribution; safety factor contours analysis was performed to assess stability of intersections.

Feb 27, 2013

As in other countries, gold mining inNew Zealand has been cyclic in character. Looking at these cycles may provide insights as to what may occur in the future and that is the basis for this study.   New Zealand has experienced four gold mining cycles. The first began in 1857, peaked in 1866 and terminated in 1890. This cycle was characterised by alluvial gold mining by individuals and small groups. This is referred to as the Labour Led cycle. The second cycle began in 1890 and reached its peak in 1906 and terminated in 1929. This cycle was characterised by the use of dredges in alluvial gold mining and was also the peak of hard rock mining. This cycle is referred to as the Technology Led Cycle. The third cycle peaked in 1940. This cycle was far less significant than the previous cycles, arising out of the depression when the unemployed embarked on small scale alluvial mining. This overprinted the continuing decline that occurred after 1906. This cycle is here referred to as the Depression Overprint Cycle.   The fourth cycle began in 1984 when production began to climb rapidly from a very low base. The growth in production was spurred by rising gold prices in the early 1980s resulting in relatively low grade gold-bearing alluvial ground becoming economic. But in the 1990s, production from the new hardrock mines in Coromandel and Otago has overshadowed alluvial production. This cycle is here referred to as the Price Led cycle.   What are the future prospects for gold in New Zealand? Although there are many factors involved, this will be largely determined by two factors; firstly, continued access to prospective gold-bearing properties, and secondly, the price of gold. Based on these factors, three future industry scenarios have been developed. The first of these is industry collapse, in which access to prospective property is constrained and prices mayor may not weaken. The second is industry consolidation, in which the available ground is accumulated by a small number of large companies who can use economies of scale to weather price fluctuations. The third scenario is industry growth, in which access to prospective ground is constrained in environmentally sensitive areas but is otherwise relatively unconstrained. Efficient companies are allowed to work in a policy and economic framework that allows them to make a major contribution to the economy.   While the industry cannot control these critical factors, a good knowledge of the positive role gold mining could play in the future economy may provide a framework by which the access issue, in particular, may be addressed from a balanced perspective.

Jan 1, 1994

By Marc Ruest

A new instrumented cable bolt has been developed by the National Institute for Occupational Safety and Health at the Spokane Research Laboratory in Spokane, WA. Although various instruments are available to measure load distribution and magnitude along a grouted cable, this instrument is innovative in that it uses strain gauges internal to the cable as load-measuring sensors. The instrument has been successfully tested in the field at FMC?s Granger Mine, Meikle Mine, and Getchell Mine. A ?split-pipe? laboratory pull test was performed on each of three cable bolts instrumented with the new device. Each test consisted of grouting a 1.83-m-long cable in two 0.91-m-long sections of schedule 80 pipe. A numerical analysis was then performed in which laboratory boundary conditions were simulated and model properties derived from textbook guidelines. The loads calculated by the model were then compared to the measured loads.

May 1, 2002

By Dominic Fragomeni, Tony DiFeo, Martyn P. Hay

"As part of an investigation into flotation modelling of industrial scale operations, Falconbridge Ltd organised a “blind” test of the SUPASIM flotation model developed by Eurus Mineral Consultants (EMC). Two sets of rougher rate data, a flowsheet and the number and size of cells in the rougher bank were supplied from a plant that had been previously surveyed in November 2003. The name of the plant, cleaner circuit details and circuit mass balance were, at the time, unknown to EMC. With only these data, flotation performance was simulated and the cleaner circuit sizedFalconbridge then compared predicted and actual float performance and mass balance and found that final concentrate recovery and mass and circuit cell volume matched closely. Simulated recoveries relative to actual were +0.4% Cu, +0.1% Ni and -0.7% Co. However, prediction of metal distribution in the internal streams making up the various circulating loads showed some sizeable errors. On this initial success, the trial was extended and additional float test data for rougher feed, cleaner feed and cleaner scavenger feed was generated on site in April 2005. Further simulations significantly improved prediction accuracy for metal distribution in the internal streams. The average difference between predicted and actual recovery for all streams was reduced from 10.5% to 0.4%. This updated simulation was then used to ascertain the impact on final recovery, grade and plant circulating loads of increasing feed tonnage from 114 to 150t/hr. A number of scenarios were simulated in order to understand the contribution that (a) the fast and slow floating fractions and (b) regrinding of cleaner tailings are making to overall plant recovery.Simulation was then used to determine how much the flotation kinetics of both rougher feed and cleaner scavenger feed have to increase in order to deliver a 1.5% improvement in overall plant recovery. These improved kinetic values provide a target for laboratory testwork and operating conditions can be varied to determine how best to achieve these new values of fraction and rate."

Jan 1, 2006

By M. R. Bitarafan

Mining method selection is the first and most important problem in mine design. In this selection some of the parameters such as geological and geotechnical properties, economic parameters and geographical factors are involved. The mining engineer must balance the input parameters and select the method that appears to be the most suitable; mining method selection is both an art and a science. Distinguishing all parameters that have an effect in mining method selection and in the determination of weighting factors of any parameter is a difficult task in an ordinary environment. This paper attempts to determine important parameters that impact on the mining method selection and demonstrates the calculation of the weighting factors for each selected parameter. This study has been directed to the research of a methodology for the selection of an optimal mining method. Different fuzzy methods are presented as an innovative tool for criteria aggregation in mining decision problems. Two methods of fuzzy multiple criteria decision making are used for selecting the optimal method with which to mine anomaly No. 3 of the Gol-Gohar iron mine. This paper finally presents a comparison of the selection methodologies and further discusses the most appropriate method investigated.

Jan 1, 2004

Gippsland has perhaps the most colourful history of exploration of any area in Australia. Hydrocarbons were first discovered in the Gippsland Basin in 1924 when a water well at Lakes Entrance encountered oil and gas shows. In 1942, the Commonwealth and Victorian Governments, in an effort to supplement a wartime oil shortage, sank a shaft near this well and produced 8000 barrels of low gravity oil up to 1956, when the field was abandoned. A total of 64 shallow wells had been drilled in the Lakes Entrance area by that time. Elsewhere in the onshore Gippsland Basin, 30 wells were drilled from 1925 to 1941. In 1960 BHP Ltd was granted an exploration permit covering much of the offshore area of the Gippsland Basin. Esso Australia Ltd joined BHP in 1964 and drilled the Barracouta-1 gas discovery in 1965. Since then over 100 exploration and appraisal wells have been drilled in the basin resulting in the discovery to date of some 3.6 billion barrels of recoverable oil and 8.8 trillion cubic feet of recoverable gas (Current remaining recoverable reserves are 906 million barrels of oil and 5.7 TCF of gas). Exploration in the offshore areas is actively continuing today and new discoveries are still being made.

Jan 1, 1992

By E. R. Maize

[In report of Investigations 3452,5/ baling with roof movement in the Crucible mine, reference was made do to the fact tart timber decay was being studied at the time of writing. This roper, written after completion of that study, describes s the work and summarizes the data obtained. When the study of roof movement as begun at Crucible two ideas were advanced on decay Of timber by some of the officials: (1) That timber decay vas unusually rapid in the Crucible mine, and (2) that timber coated with rock dust decayed more rapidly than those teat were not so coated. There yore no records by which these ideas could be tested, and it appeared wise to make a limited; investigation particularly of the effect of rock dusting on decay, as rock dusting is an important means of preventing coal-dust explosions.]

Jan 1, 1940

By James A. Schneider

Selection of resistance factors within Load and Resistance Factor Design (LRFD) requires accurate assessment of uncertainty and bias in load and resistance. To date, many studies have focused predominantly on the assessment of uncertainty, while the potential for bias appears to be the dominant factor in design for consistent levels of reliability under a variety of situations. Bias comes from factors which have been ignored in design method development but significantly influence resistance when extrapolating from conditions typical of databases to those at a given site. Uncertainty and bias can be quantified as the coefficient of variation and mean (or median) of a data set, respectively, for statistical assessment of load and resistance factors. This paper first presents an overview of statistical assessment of resistance factors, and then discusses factors that influence uncertainty and bias when extrapolating to situations outside of conventional databases of driven pile axial capacity.

Jan 1, 2009

By Y. Faitakis, G. J. Powley, C. Mackenzie

"In the mining industry, maintenance is a key priority and a major expense. As mining equipment is large in scale, high in cost, and difficult to tow, it is key to maintain equipment on a timely basis. Breakdown in -40°C or away from a maintenance station can cause great costs in repair as well as lost production.It has often been quoted that about 5% of the North American production is lost every year due to unscheduled downtime. About onethird of the downtime is attributable to equipment failures. From this fact alone, it is evident that proper equipment maintenance can add tremendous value to the bottom line. As organizations and equipment become more sophisticated, maintenance expectations begin to evolve. This evolution can be traced through three generations:The first generation (1940–1950): A “fix it when it broke” attitude prevailed during that time, mostly because the industry was not highly automated.The second generation (1950–1980): Automation and mechanization of the industry increased. The industry now was dependent on machines performing the majority of the work. Preventive maintenance (overhauls at fixed intervals) appeared as the first maintenance practice. Maintenance costs started to rise rapidly.The third generation (1980–): Assets become more complex. Failure modes are better understood and computerized maintenance methods are becoming mainstream. Condition-based maintenance (CBM) starts to gain ground as a cost-effective alternative, replacing the fixed interval overhauls. As corporations amalgamate and market places are becoming more complex multi-functional teams become responsible for maintenance.At the beginning of the third generation, maintenance organizations began embracing technology to provide efficiencies in maintenance management. A strong push was made to transition from reactive maintenance to proactive maintenance. At the forefront of this push is condition-based maintenance."

Jan 1, 2004

By Y. Faitakis

In the mining industry, maintenance is a key priority and a major expense. As mining equipment is large in scale, high in cost, and difficult to tow, it is key to maintain equipment on a timely basis. Breakdown in -40°C or away from a maintenance station can cause great costs in repair as well as lost production. It has often been quoted that about 5% of the North American production is lost every year due to unscheduled downtime. About one-third of the downtime is attributable to equipment failures. From this fact alone, it is evident that proper equipment maintenance can add tremendous value to the bottom line. As organizations and equipment become more sophisticated, maintenance expectations begin to evolve. This evolution can be traced through three generations: The first generation (1940?1950): A ?fix it when it broke? attitude prevailed during that time, mostly because the industry was not highly automated. The second generation (1950?1980): Automation and mechanization of the industry increased. The industry now was dependent on machines performing the majority of the work. Preventive maintenance (overhauls at fixed intervals) appeared as the first maintenance practice. Maintenance costs started to rise rapidly. The third generation (1980?): Assets become more complex. Failure modes are better understood and computerized maintenance methods are becoming mainstream. Condition-based maintenance (CBM) starts to gain ground as a cost-effective alternative, replacing the fixed interval overhauls. As corporations amalgamate and market places are becoming more complex, multi-functional teams become responsible for maintenance. At the beginning of the third generation, maintenance organizations began embracing technology to provide efficiencies in maintenance management. A strong push was made to transition from reactive maintenance to proactive maintenance. At the forefront of this push is condition-based maintenance.

Jan 1, 2004

By R. J. Snoby, J. K. Alderman

Although once fairly common as a method of coal preparation in the United States, dry cleaning has virtually disappeared. Low-rank coals used for power generation in the western U.S. are suited for dry cleaning because water has limited availability in the American west. Surface moisture added through wet cleaning offsets energy improvement achieved from mineral matter reduction. Information will be presented on the commercial application of dry cleaning in North America, the historic development and current status of the technology, and the introduction of modern processing theory and control technology to the design of dry cleaning equipment and plants.

Jan 1, 2001

By Dennis R. Dolinar

Controlling coal mine roofs subjected to high-horizontal stress conditions has always been difficult and uncertain. Traditional supports such as wooden cribs and posts, concrete donut cribs, and standing supports collapse and fail when the roof and floor move horizontally as mining progresses. The former U.S. Bureau of Mines (USBM) (currently the U.S. Department of Energy (DOE)), in cooperation with Western Fuels-Utah, Incorporated, conducted research to provide an alternative for traditional secondary support systems in a 3-entry gate road system subjected to high horizontal movements. The support system used in several other coal mine operations, consisted of internal high-strength galvanized resin-grouted cable supports. The system virtually eliminates the necessity for external crib, timber, or concrete supports. The support system consisted of 2.4 m (8 ft) full-column resin grouted bolts and 4.8 m (16 ft) long cable supports installed in conjunction with wire mesh and "Monster-Mats." Cable loading and roof deformations were monitored to evaluate the behavior of the immediate and main roofs during first and second panel extractions. Additionally, cable trusses were installed on the longwall headgate to protect the coal conveyance system from roof and pillar falls created by the formation of cutters and gutters. The test results indicated that the designed support system successfully maintained the roof during the extraction of two longwall panels and dramatically reduced the cost of secondary support. This paper describes the theory of high-horizontal roof movements, the advantages of vertical cable supports and cable trusses, and presents the roof and cable measurements made to assess the support performance during longwall retreat mining.

Jan 1, 1996

By N F. Fraser

Seafield ResourcesÆ offshore placer gold project covers c.10 000 km2 of the Westland continental shelf between Karamea and Jacksons Head, from the low-water mark out to the 120 m isobath (Figure 1). De Beers Marine (DBM), the marine diamond subsidiary of the De Beers Group, will carry out the project with geological input from Placer Solutions (PSL). DBM bring a wealth of expertise and advanced technology to the project, developed during their 20 years of exploration and mining of diamond placers on the Atlantic shelf of southern Africa, where they have also maintained a strong environmental record.   Glaciers and high-energy rivers emanating from the Southern Alps have transported substantial quantities of gold bearing sediment onto the Westland continental shelf during Pleistocene glacial periods, when fluvial gradients are steeper, sea level is c.120 m lower and most of the shelf is exposed. These deposits have been extensively reworked by the advancing shoreline and marine processes during interglacial marine transgressions, and most are now buried beneath a variable thickness of post-glacial marine sand, silt and mud. DBM will use side-scan sonar, swath bathymetry, Chirp sub-bottom profiling and a magnetometer to map the sea floor and sediments beneath it, guided by PSLÆs predictive geological model for gold distribution on the shelf. Cactus grab and vibro-core samples will be used to ground-truth the geophysics and provide stratigraphic, sedimentological and gold distribution data for incorporation into the geological model. Areas of interest identified by the geophysical and sampling data will be investigated in more detail, with closer-spaced geophysics and Mega-drill sampling to a maximum sediment depth of 20 m beneath the sea floor. These samples will be raised to the ship with air-lift suction technology, processed onboard for mineralogical and grade information, then returned to the sea floor. Environmental baseline data will be gathered from Van Veen grab samples of seabed sediment taken throughout the area during the prospecting programme.

Jan 1, 2005

By J. H. Schubert

Mintek has developed a level-control system, FloatStar, for use in flotation circuits. Instead of controlling levels separately in each unit, the controller monitors levels throughout the circuit. It can thus take account of the global effects of any control actions. A collaborative project between Mintek and Karee Platinum Mine has begun, looking at aspects of flotation control. Karee plant was chosen as the site for this work because of its good existing control infrastructure. As a first stage of the joint project FloatStar was installed. Analysis during installation and running of the FloatStar has illustrated how circuit design and measurement technology can affect process control. Severe interactions have been observed in the levels of adjacent banks. The effect of FloatStar on flotation performance was tested and an improvement in level control was observed. Similarly, a faster settling of the plant after start-ups was observed. Both the improved level control and faster start-ups lead to higher platinum recovery.

Jan 1, 1999

By Jose R. Hernandez-Aguilar

"This paper revisits data from previous studies (Hernandez-Aguilar, 2010, 2011) to better understand the link between metallurgical performance and pulp-zone bubble size (Db) in full-scale flotation systems. To obtain conclusive evidence exclusively related to the pulp (collection) zone, the experiments were designed in those studies to control both the bubble generation method and the influence of the froth zone on the overall metallurgical performance in industrial and pilot-scale flotation columns. The results showed that Db had a significantly larger effect on the recovery of coarse (> 75 µm) particles compared to the effect on fine (< 16 µm) particles. It was observed that, under certain practical conditions, the recovery of coarse particles was close to zero, but dramatically increased by reducing Db. The evidence revealed an inconsistency of the so-called “k–PSb” model: at constant froth recovery, increasing Sb (i.e., the rate of production of bubble surface area) did not always result in higher recovery, as the model predicts. Furthermore, the evidence showed that it was not the Sb but the bubble size that proved to be a more reliable link to metallurgical performance. Only in the presence of suboptimal-sized bubbles (Db > 2 mm) did the froth zone seem to play a role, possibly reflecting an effect of pulp-zone bubble size on froth stability. However, the role of the froth was negligible for Db < 1 mm.High-resolution mineralogical measurements performed on process samples revealed that, regardless of the degree of liberation, Db had a profound effect on the shape of the recovery-by-size curve, particularly the coarse particle size fraction. This observation is key as it implies that the loss in efficiency associated with floating the coarse particles was not the result of neither particle hydrophobicity nor particle detachment issues: even in the absence of mechanical agitation, liberated and highly hydrophobic coarse particles did not seem to attach to the bubbles in the first place, an inefficiency that was corrected by reducing bubble size. The data analysis suggests that poor attachment efficiency can practically affect both fine and coarse particles but via different mechanisms which are discussed here."

Jan 1, 2017

By M. I. Brasileiro

The use of experimental design to the study of mixtures has found a wide range of applications, even in laboratory scale or in industrial development works. The aim of this work was to maximize granite sawing waste content in ceramic tiles using experimental design. Based on the raw materials, specific formulations were developed using the experimental design. The raw materials were mixed and sample bodies were produced using uniaxial pressing. The sample bodies were fired and characterized in terms of water absorption, mechanical strength and X-ray diffraction. Regression models were adjusted relating the water absorption and mechanical strength with the amount of raw materials. The results showed that the methodology of experimental design maximize the content of wastes incorporated in ceramic tiles formulations. Formulations containing up to 40% of waste can be used for the production of ceramic tiles.

Jan 1, 2009

By A Hekmat, M Fuentes, F Armijo, R Castro, F Rodriguez

Interpreting the behaviour of granular material and simulating the gravity flow are complex processes in cave mining methods. In this paper, a new Cellular Automata based code – FlowSim – is used as a flexible tool to simulate the gravity flow and subsidence zone in block caving mines. The simulation results of the gravity flow model in a real scaled mine as well as a scaled physical model are presented in this study. In this research, FlowSim was first calibrated at a mine, based on the information obtained from smart markers. These markers were located in five points of the second block of the Esmeralda sector, at the El Teniente copper mine, Codelco, Chile. The calibration was performed with the objective of reducing the discrepancy between the extraction zone defined by markers in the mine and FlowSim outputs. Regarding the mine data, FlowSim was calibrated with the minimum error of ten per cent between real and simulated results. Then, the calibrated FlowSim model was used to simulate the flow behaviour of a scaled physical model. This model was constructed in the Block Caving Laboratory of the University of Chile to scale down the same block of the Esmeralda sector with the scale of 1:200. The FlowSim outputs show that this software enables the simulation of the gravity flow of the scaled model with less than five per cent discrepancy in grade and dilution entry to the drawpoints. According to the obtained results, this paper recommends FlowSim as an appropriate tool which can improve the design of production level in block caving and can enhance the prediction of the subsidence zone and dilution entry.CITATION:Castro, R, Hekmat, A, Fuentes, M, Armijo, F and Rodriguez, F, 2016. FlowSim – a versatile flow simulation tool to quantify extraction and design alternatives for block caving, in Proceedings Seventh International Conference and Exhibition on Mass Mining (MassMin 2016), pp 645–652 (The Australasian Institute of Mining and Metallurgy: Melbourne).

May 9, 2016

By R. B. Booth, W. F. Linke

The continuous interest of the American Cyanamid Company in producing superior polymeric flocculants and dispersants for the mining industry has resulted in a broad, general study of the physical chemistry of mineral-polymer systems. The aim of this program has been to establish a sound theoretical basis on which practical problems could be analyzed, and which would lead to the development of new and improved flocculating and dispersing agents. This study has required the cooperative effort of physical, analytical, and polymer chemists, as well as engineers familiar with mining problems. This report summarizes the results of some of our investigations over the past two years. Further studies along these and other lines are in progress and will be reported in the near future. MEASURING FLOCCULATION IN THE LABORATORY The foremost problem in studying flocculation in the laboratory has been to establish reproducible methods of comparing and evaluating the effectiveness of various reagents. Previous workers have measured settling -rates 1,8 filtration rates2 , and sediment volumes3 . Direct observations of the floc size, or the clarity of the supernatant liquid, can also be used. AIthough these parameters are all interrelated, they do not always measure the same physical properties of the suspensions, and hence are not equally useful or reliable. Whatever the method of comparison selected. the testing procedure must be carefully defined and standardized. The results must be satisfactorily reproducible, and for practical purposes must accurately reflect the results obtained in mill use. In order to compare the data obtained in these various types of test, the physical nature of the flocculated system must be examined in some detail. In an unflocculated system, only the larger or denser particles (the "sands") settle at a useful rate. The sediment can be observed to build up from the bottom of the cylinder, while the fines remain in suspension for a long period of time. If the system is not disturbed, even the microscopic fines can be seen to classify themselves as they fall over a period of days. When a high molecular weight polymer is added to the system, the colloidal particles are tied together randomly in flocs, and the sands and fines settle together. The line of demarkation between solid and liquid phases moves downward from the top of the cylinder. The settling rate of the flocs depends on their size and on their effective density. For a given mineral sample, the larger the flocs, the faster the settling. Since the large flocs cannot pack as closely as small ones, they produce a sediment with more pore spaces, and hence of larger volume, which is easier to pump and to agitate. Although it is nearly impossible to measure individual floc sizes, it is relatively easy to determine settling rates and sediment volumes. Of these two, the settling rate is simpIer to measure, and hence generally more practical. Reproducible sediment volumes can only be obtained by allowing the solids to pack for a rather long period of time (e.g., hours to days), and although the volumes thus observed are directly related to the degree of flocculation (average floc size), it is often impractical to wait so long for the results.

Jan 1, 1961

Paper presented at the IAGOD international symposium on mineralization related to mafic and ultramafic rocks held in Orleans, France, 1-3 September 1993. Phases of the Ir sub-group are present either as inclusions in or interstitially to chromite and are unaffected by (or only altered in situ during) hydrothermal processes. By contrast, almost all phases of the Pd sub-group are located in a secondary textural setting - in fractures in chromite or between fragments of brecciated chromite. The different settings indicate that the mineralisation episodes were associated with magmatic and hydrothermal stages respectively. The presence of an included Pd-group mineral in a hydrothermally Pd-group mineralised chromitite sample suggests that the hydrothermal Pd-group minerals were at least partly derived by remobilisation from magmatic deposits in the chromitite pod. These observations indicate that Pd-group elements are more mobile than Ir-group elements under hydrothermal conditions and point to the potential for hydrothermal platinum and palladium mineralisation in favourable environments

Apr 1, 1994

By Charle Pelley, Kosta Fytas, Peter Calder

"A computer model has been developed which is capable of evaluating alternative production schedules (long- and short range) of an open pit mine once the ultimate pit limits have been defined. The model user can choose the best alternative according to cash flow considerations and the particular mining conditions. The model involves a combination of simulation for long range production scheduling and linear programming for short range scheduling. IntroductionMining of mineral deposits in such a fashion so that, at depletion the maximum possible profit is realized, has been a very old problem of the mineral industry. In the past, when high-grade ore reserves were readily available and profits were very high, no question of optimum profitability was raised.However, recently the complexity of both open pit and underground mining has been increased tremendously. The additional problems of continuously rising production costs on one hand and of depressed metal prices worldwide on the other, have been demanding optimum profitabiliy from every mining venture .Particularly in the area of open pit mining, the sequence of ore and waste extraction has now become very important since the size of the pits and of the mining equipment has increased enormously. The process of open pit mining involves a continuous change of the open pit surface and there are many feasible plans of extraction which lead to the same final pit shape (ultimate pit limits). Each of these feasible plans generates a production schedule over the entire mine life, which mayor may not be optimal."

Jan 1, 1987