Background

Modern society requires a sustained supply of minerals for its growing and nutriment. Industrial minerals such as limestone, salt, silicon oxide, clay, gypsum etc. are critical to the societal and economic growing. These minerals provide indispensable natural stuffs which underpin the fabrication and building industry. Cement natural stuffs are amongst the major industrial minerals including limestone, shale, slate, clay and marl etc. Cement production involves processing of these selected mineral natural stuffs to bring forth a man-made mineral mixture ( cinder ) that can be ground to pulverize transporting the specified chemical composing and physical belongingss of cement. A simple layout of the cement fabrication operation consists of four stairss including ( Austin, 1984 ) :

Mining of natural stuffs from the quarry/quarries,

Developing a natural mix consisting of natural stuffs from the prey and additives from the market,

Processing ( firing ) of the natural mix in a cement kiln to bring forth a merchandise called “ cinder ” , and,

Crunching of the cinder for distribution in different signifiers to the clients as cement.

Therefore, the handiness of natural stuffs incorporating needed measure and quality of oxides of Ca, silicon oxide, aluminium and Fe i.e. CaO, SiO2, Al2O3, Fe2O3 is the stipulation for the origin of a cement fabrication operation. Limestone quarries chiefly lend these desirable components for the development of natural mix, nevertheless, in order to run into the stringent quality demands of the cement works, it is compulsory to intermix high and low-grade natural stuffs in the prey and, if required, with additives such as sandstone, wing ash, Fe ore, clay, slate rock, etc. normally purchased from the market ( Kathal, 1999 ) .

Consequently, an optimum prey production program must guarantee the supply of natural stuffs at least possible cost, such that, the limestone mined in a given period meets the measure and quality/raw commixture restraints and the needed per centum content of the critical chemical components in the natural stuffs is achieved ( Rehman, 2010 ; Asad, 2008 ) . This indirectly guarantees a cost decrease for the purchase of expensive additives from the market.

Therefore, a successful cement fabrication operation depends upon natural stuffs be aftering and control and such control must needfully get down at the really beginning of the production i.e. in the limestone preies. The natural stuff resources are an of import plus and must be rationally, economically and expeditiously managed, exploited, and used ( Rehman, 2005 ) . Planing and scheduling are important to cement quarry operations, merely as to other excavation and technology spheres. Mining companies throughout the universe have recognized that considerable additions are accomplishable through careful planning of mining operations. Keeping the significance of this facet in position, this research embarks upon the development of optimum long and short term prey production programs.

Quarry Production Planning

Exploratory boring, function, and geological reading set up the size, form, deepness, and geology of the limestone militias. Geostatistical mold of this information dictates the division of limestone militias into a figure of mineable blocks, where each block is assigned both quantitative and qualitative informations in footings of available tunnage and per centum content of possible chemical components including SiO2, Al2O3, Fe2O3, CaO, MgO, SO3, K2O, Na2O, TiO2, P2O5, and Cl, etc. ( Ramazan, 2007 ) . Therefore, a 3-dimensional block theoretical account as presented in Figure 1.1 and 1.2 becomes an input to the prey planning activity, which may be divided into two stages, i.e. long and short term planning.

Long term programs crossing over the life of an operation ascertain the rightness of the natural resources for cement production, extent of the prey, appraisal of prey life-time, contractual demands for sustained supply of additives from the local market and overall profitableness of the operation. Long term prey planning and design undertakings are typically in focal point when planing the cement works. These are non of the nature that one needs to animate them every twenty-four hours, hebdomad, or month. These undertakings are reassessed with long term intervals or in instance of particular fortunes. These fortunes may include alteration in production capacity, enlargements through leasing and land acquisition, alteration in direction, replacing of machinery, and betterment in engineering, etc.

Once a determination is made for the capital investing, the following measure is to develop a series of short term production programs embracing operational schemes to accomplish long term aims, hence, the optimality of short term programs may non be compromised. Short term be aftering includes informations import from established optimum long term programs, dynamic appraisal of prey beginning stuff chemical science, natural mix optimisation to run into current reserve mark chemical science i.e. rate control of the incoming stuff watercourses, cost minimisation to guarantee lowest possible runing cost of feeding reserves, truck/dumper fleet direction, informations storage, visual image and coverage ( Pedersen, 2006 ) . Therefore, prey planning, design, and operations are considered an incorporate portion of cement fabrication procedure. Quarry operations constitute the procedure sequence concerned with the transmutation of heterogenous prey stuffs into a homogenous natural blend ( Mortensen, 1987 ) .

Optimization techniques have been successfully implemented to work out jobs originating in prey planning and production programming as traditionally these determinations were resolved through intuition, experience, and judgements. Computational tools have been developed to back up these procedures. Significant work has been done in this respect. This includes literature on the issues of optimum planning and programming and algorithms developed by assorted research workers ( Gershon ( Gershon, 1982 ; Gershon, 1983 ; Gershon, 1987 ) , Mann & A ; Wilke ( Mann, 1992 ) , Wang & A ; Sevim ( Wang, 1995 ) , Dagdelen & A ; Johnson ( Dagdelen, 1987 ) , Dagdelen & A ; Francois ( Dagdelen, 1982 ) , Francois-Bongarcon & A ; Guibal ( Francois – Bongarcon, 1984 ) , Chanda ( Chanda, 1990 ) , Tolwinski & A ; Underwood ( Tolwinski, 1998 ; Tolwinski, 1996 ) , Fytas et Al. ( Fytas, 1987 ) , Caccetta & A ; Hill ( Caccetta, 2003 ) , Kumral & A ; Dowd ( Kumral, 2002 ) , Denby & A ; Schofield ( Denby, 1994 ) , Denby et Al. ( Denby, 1998 ) , Ramzan ( Ramazan, 2001 ) , Asad ( Asad, 2001 ) , Lindon et Al. ( Lindon, 2005 ) , Onurgil ( Onurgil, 2005 ) , Goodwin et Al. ( Goodwin, 2006 ) . Similarly a assortment of commercial package bundles such as GemcomA© , MineSightA© , SURPACA© , VulcanA© , QSOA© , FLS – CombiLogic Systems and DatamineA© etc. are based on these algorithms and are used by mine contrivers ( Mortensen, 1987 ) .

Search for comprehensive mathematical theoretical accounts that would optimally work out production scheduling jobs is an active country of research for mathematicians, computing machine scientists, and excavation applied scientists. Artificial intelligence techniques including familial algorithms, unreal nervous webs have besides been pragmatically applied. These algorithms and their alterations have been successfully used. The focal point of this research is to develop a familial algorithm based long term production program and a assorted whole number additive scheduling ( MILP ) based short term production agenda for cement prey operations.

Preliminaries

Presently, the cement industry patterns a manual interaction i.e. test and mistake attack utilizing MS ExcelA© spreadsheets to carry through prey production planning and natural stuffs intermixing ( Asad, 2001 ) . However, this pattern fails to measure alternate options and it besides ignores the effects of block precedency and sequencing that leads to a deficiency of coordination among assorted subdivisions of the operation ( Asad, 2008 ) .

The demand to be after and run limestone preies with optimum production planning has been acknowledged. However, with few exclusions, the greater portion of surveies on production planning and blending of natural stuffs have been conducted with a focal point on unfastened cavity excavation operations, chiefly applied to the excavation of metallic ores. The most of import grounds for this inclination include:

The big magnitude and processing complexnesss associating an unfastened cavity mining operations compared to somewhat simpler characteristics of a prey operation.

A instead complex scenario of grade distribution in unfastened cavity compared to a reasonably unvarying composing distribution of cement natural stuff preies.

Large volumes of ore and waste to manage in unfastened cavity mines as compared to volume of stuff produced in quarry operations.

Significant impacts of altering economic conditions and metal monetary values on cut off class and unfastened cavity planning and design.

A brief description of few parts is presented in the undermentioned subdivisions.

Nested Lerchs and Grossmann ‘s ( LG ) Algorithm

Lerchs-Grossmann ( Lerchs, 1965 ) introduced the construct of “ parametric analysis ” , in which the development of a cavity is characterized by the gradual modificaA­tion of one or more cardinal parametric quantities. By this manner an extraction sequence was produced which would maximise the net hard currency flow with regard to the entire volume mined. LG algorithm selects a parametric quantity for cut downing the economic value of each block in the theoretical account. If this parametric quantity is zero, the normal ultimate cavity is produced. As it is increasingly increased past the critical values, the ultimate cavity contour becomes smaller. The terminal consequence is a series of nested cavities which may be used to deA­rive a suited extraction sequence.

The nested cavities are produced if there are non excessively many mutualities beA­tween the ore and waste parts in the sedimentation and there is suffiA­cient fluctuation in the economic values of the blocks. Unfortunately, if these conditions are merely partly satisfied, the leap between back-to-back nested cavities may be so big that it hampers the effectual application of the nested cavities to the extracA­tion sequencing procedure. This job is normally referred to as the “ spread job ” . Indeed, under the “ right ” conditions, it is even possible to leap from no cavity to the complete ultimate cavity in one increase.

Parameterization Technique

Parameterization technique is used in the coevals of nested cavities. Francois-Bongarcon & A ; Guibal ( Francois – Bongarcon, 1984 ) , Dagdelen & A ; Johnson ( Dagdelen, 1987 ) , Whittle ( Whittle, 1988 ) and Coleou ( Coleou, 1988 ) expanded the thought of parameterization to include proficient every bit good as economic parametric quantities of the sedimentation. The parameterization technique consists of repeated usage of one of the ultimate cavity bound algorithms on different economical block theoretical accounts. This process is based on the fact that altering block values will ensue in different cavity sizes. Therefore, increasing block values consistently generates nested pit lineations. The most common and simple manner to alter block values is by changing the concluding merchandise merchandising monetary value. This process generates several different cavity sizes. The mine planning applied scientist may utilize these cavity lineations as a usher for finding of the cavity extraction sequences.

Simulation and Linear Programming

Fytes et Al. ( Fytas, 1987 ) developed a computing machine theoretical account PITSED for unfastened cavity long and short term production programming. The theoretical account uses simulation technique for long term production programming and consists of three theoretical accounts viz. simulation theoretical account, smoothing theoretical account and hard currency flow analysis theoretical account. Similarly a additive scheduling theoretical account is used for short term production programming. PITSHED generates alternate schemes accurately and helps in determination devising procedure. The chief drawback of PITSHED is that the long term agendas generated are non guaranteed to be optimum.

Dynamic Programming, Artificial Intelligence and Heuristic Search Algorithms

Tolwinski and Underwood ( Tolwinski, 1998 ; Tolwinski, 1996 ) developed algorithms for the finding of a production agenda for an unfastened cavity mine. The algorithm is based on thoughts from dynamic scheduling, unreal intelligence and heuristics hunt regulations. The algorithm is developed by uniting constructs from both stochastic optimisation and unreal nervous webs. The development of the mine is modeled over clip as a consecutive optimisation job. The purpose is to happen the highest-valued way ( defined as a patterned advance from one possible cavity to the following ) through a sedimentation, in the procedure bring forthing incremental cavity deA­signs and an extraction agenda at the same time. By trying a little, but hopefully representative, part of all of the possible waies through a sedimentation, the algorithm attempts to larn those features of a way which cause it to bring forth a high value, and, conversely, those which cause it to bring forth a low value.

The passage from one cavity to the following in a way ( referred as a “ province alteration ” ) is governed by a chance disA­tribution and based on the figure of times that a peculiar province has appeared in a extremely valued way. The clip value of money, incline limitations, minimal equipment runing breadths and demand for a unvarying flow of natural stuff to the factory are all considered by the algorithm. The working breadth is defined utilizing the bead cut construct. Here a block is removed as a bead cut if all of its surrounding blocks are at the same degree. When a bead cut is made, there is sufficient working infinite. The angle of disposition of any cone is chosen to match to the incline angle for the sides of the cavity.

Even though this algorithm attempts to include some parametric quantities vital to open cavity design farther work is required to better it. The algorithm suffers from certain restrictions including:

Unable to specify certain parametric quantities in a mode applicable to existent excavation state of affairs.

Definition of working breadth utilizing bead cut is incompatible with excavation pattern. It does non integrate bench tallness which is a map of the geology and the bing excavation equipment and so is seldom changed during the life of the cavity unless there is a important alteration in economic parametric quantities which accordingly alters recoverability standards.

Improvement in techniques like aggregation/disaggregation strategies to rush up larning every bit good as more sophisticated information constructions to do better usage of computing machine memory.

Other characteristics of scheduling job may hold to be incorporated into the theoretical account to suit possible users.

It suffers to some extent from combinative detonation effects.

Familial Algorithms for Open Pit Design and Production Scheduling

The usage of familial algorithms ( GAs ) in the design and programming of unfastened cavity mines was foremost proposed by Denby and Schofield ( Denby, 1994 ; Denby, 1998 ) . This can be considered as a pioneering work in turn toing the cavity planning and scheduling utilizing one of the promising evolutionary calculation and hunt techniques. They presented a technique which at the same time provides a combined cavity bound and extraction agenda that aims to maximise the net nowadays value ( NPV ) of the ore in the cavity. In this algorithm, the value of a cavity is calculated based on the value of each block when it is mined.

Familial algorithms mimic the operations of genetic sciences and natuA­ral choice in their hunt for the optimal solution. They begin their hunt with a population of random solutions and germinate this population over a series of coevalss by using chance techniques and familial operators to each member of the populaA­tion. To come on from one population to the following, capable to their current fittingness, members of the population are reproduced, crossed over, mutated and/or inverted. When using GAs ( and other stochastic optimisation techniques, such as fake tempering ) to the mine design-scheduling job, careful consideration must be given to the followers:

Formulation of the job,

Design of the cross-over and mutant operators, and

Standards for measuring when the optimisation should be terminated.

The optimisation processs are as followers:

Coevals of random cavity population ( figure of cavities ) with a higher mean value.

Appraisal of cavity fittingness value whose computation is independent of the chief optimisation procedure for each of the agendas in the population.

Reproduction of pit population during which each agenda either survives to the following coevals or is removed wholly. These processs produce a population with higher norm value.

Crossover of cavities during which selected agendas are indiscriminately combined in braces on a probabilistic footing. Using crossing over operator to a new population produces a new population of cavities uniting countries ( characteristics ) of bing cavities to make both higher and lower value cavities.

Mutant and standardization of cavities. Mutant is performed on probabilistic footing, a little per centum of the cells in the agenda being modified in a random mode to assist keep familial diverseness and prevents the system from meeting to a false optimum. Normalizing maps are so applied to new population. This changes the cavities generated into executable scheduled cavities by the application of a figure of constraint maps.

After repeating through a figure of coevalss an optimal cavity and agenda is produced at the same time.

The chief restrictions of utilizing GAs for cavity design and programming are as follows:

The system slows down bit by bit as the size of the job additions, but if this can be minimized for larger jobs, a practical system is executable.

If the job can be formulated in such a manner that the crossA­over and mutant operators can be applied without bring forthing solutions that violate critical restraints, the overall efficiency of the optimisation procedure can be improved well.

As each possible solution is checked for restraint misdemeanor and, if necessary, corrected and/or penalized, much of the optimisation clip is wasted bring forthing and measuring impracticable solutions. Scaling-up the system to get by with realistic jobs requires farther probe.

Ant Colony Optimization for Long Term Open Pit Planning

Sattarvand ( 2009 ) developed a mataheuristic algorithm based on the Ant Colony Optimization ( ACO ) for long term unfastened cavity planning. Main characteristics of utilizing ACO are ;

The ACO attack is able to see non merely simple nonsubjective maps like maximising NPV, but it can manage complex nonsubjective maps expeditiously.

The algorithm generates agendas irrespective of the nonsubjective map and the values of the generated agendas are calculated harmonizing to any defined mark.

Thousands of mine agendas are indiscriminately created and can easy pattern uncertainness related to the features of the blocks bring forthing agendas based on a series of the random variables alternatively of deterministic block values.

Major drawbacks of utilizing this method, which have to be considered, are:

Application of ACO and its discrepancies is really hard, or even impossible, for big block theoretical accounts, because of being to a great extent memory intensive.

The procedure is non mathematically proven to ever make the best agenda.

The efficiency of ACO algorithm is extremely dependent on its parametric quantities and there is no other manner except a hit and test attack to happen the best combination of parametric quantities.

Some discrepancies of ACO like Ant System and Elitist Ant System can non be applied for a existent sedimentation in pattern.

ACO was implemented on two dimensional instances and a 1:1 inclines but its application in a existent excavation instance was non tested.

Optimum Production Scheduling Using Receding Horizon Control

Goodwin et Al. ( Goodwin, 2006 ) presented an alternate preparation of mine planning job utilizing withdrawing skyline control. This scheme has been successful in procedure control applications. This alternate preparation has improved computational processs and incorporates many practical characteristics of an existent mine planning job. The basic thought behind withdrawing horizon optimisation is that one solves a fixed skyline optimisation job to calculate a sequence of predicted inputs over some anticipation skyline ( state T clip stairss ) but one merely implements ( or shops ) the first measure. Then clip is advanced one measure and the procedure is repeated.

Goodwin et Al. used non unvarying quantisation and collection by agencies of footing maps ( with matching alterations in the excavation capacity restraint ) , to give a simplified solution. Different trials on a practical mine planning job, utilizing a different figure of units of collection and tantamount excavation conditions, resulted in a better NPV. The preparation was advantageous in footings of penetration and computational complexness.

Quarry Sequencing and Scheduling Using Mathematical Programming

Asad ( Asad, 2001 ) used assorted whole number additive scheduling ( MILP ) to bring forth optimal production agendas for cement prey operations. He developed quarry sequencing techniques to bring forth executable one-year excavation programs and an MILP blending optimisation theoretical account is used to choose the best amongst these programs. An execution of sequencing and intermixing optimisation theoretical accounts to an existent prey block theoretical account has shown promising consequences ( Asad, 2001 ) . The theoretical account is capable of work outing for 1000s of executable excavation programs for both long and short term production programming of prey operations.

Major restrictions of this attack include separation of the block sequencing and programming, sequencing techniques does non vouch an thorough hunt of all the executable excavation programs, and inability to bring forth synchronized long and short term production agendas. Sequencing techniques can be made more flexible by leting the choice of neighbouring blocks at random without a limitation of predefined way.

Multi-Objective Simulated Annealing

Kumral and Dowd ( Kumral, 2002 ) used Multiple Objective Simulated Annealing ( MOSA ) for short term production programming of industrial minerals. The aim of short term production programming of industrial minerals was to optimise the quarry-blend-stockpiling procedure. The methodological analysis for quarry optimisation and optimum programming was divided in three phases, including:

The quarry-blend bound was determined by a combination of additive scheduling and the Lerchs-Grossman algorithm.

A sub-optimal production agenda was obtained by any conventional or heuristic method capable of giving a sensible consequence in an acceptable clip.

MOSA was used to better the sub-optimal agenda to an optimum or near-optimal agenda.

MOSA can be applied to the end product from any traditional method to better consequences. A major restriction of utilizing fake tempering is its inability to be used as a exclusive method of optimisation for quarry-blend-stockpiling optimisation jobs. The ground of this inability is the huge solution infinite.

Familial Algorithms for Short Term Production Scheduling

The flexibleness offered by GAs is its major characteristic. Familial algorithms have besides been used for the short term production programming by some research workers ( Samanta, 2005 ; Ito, 2003 ) .

Samanta et. Al. ( Samanta, 2005 ) states that the optimum agendas generated by elegant mathematical techniques become impracticable when they are turned into existent applications that violate many operational restraints and familial algorithms offer the potency for a matter-of-fact attack to the solution of bring forthing practical executable agendas. Samanta et. Al. generated a class control program ( short term production agendas ) utilizing GAs. The consequence was a figure of sub-optimal agendas in a speedy clip period. This gives the mine direction a flexibleness to make up one’s mind over which agenda will be the best suited for their practical application. The major defect of this attack was its really limited range and inability to undertake the multi aim and multi constrained existent life excavation operations. The aim of seeking for optimal agendas was grade control without taking into consideration the cost of excavation.

Ito and Nishiyama ( Ito, 2003 ) used familial algorithms and dynamic scheduling for production programming of limestone prey operations. The aims were to maximise the life of prey and to minimise the per centum content of unwanted chemical constituents in limestone taking into consideration the block precedency restraint. The consequences showed that GAs can seek optimal agendas for big graduated table excavation, while dynamic scheduling method would be effectual in short term scheduling. This attack besides suffers from the drawbacks of holding a limited range and inability to manage multiple aims and restraints.

Applicability of Open Pit Production Scheduling to Quarry Operationss

In unfastened cavity excavation operations for metallic ores, based on its class, i.e. metal content, an economic value is assigned to each block. A cutoff class determined through the economic parametric quantities including monetary value of metal, operating costs ( excavation, processing, and refinement costs ) , and metallurgical recovery discriminates between the ore, i.e. valuable stuff and the waste known as worthless stuff ( Dagdelen, 200 ; Asad, 2007 ) . Therefore, it is the standard that assigns a positive or negative economic value for designation as an ore or waste block, severally. However, as explained earlier, in cement prey operations limestone militias are divided into a figure of mineable blocks and each block is assigned a qualitative index in footings of per centum content of each of the critical chemical components required for cement production. Therefore, the solution to the cement prey production planning job is different from metallic ores due to the difference in block theoretical accounts i.e. the basic input to the production scheduling algorithms or preparations ( Asad, 2001 ; Srinivasan, 1996 ) .

The old surveies related to the production planning of unfastened cavity excavation operations could be applicable to the production planning of cement prey operations merely if the block theoretical account is translated in footings of economic values. However, the undermentioned grounds ascertain that this transition of the block values is impractical ( Asad, 2008 ; Chanda, 1995 ) :

The refined metals recovered from valuable stuff are sold as a trade good in the metal exchanges. Therefore, the monetary value of a metal becomes a footing for the block theoretical account of an unfastened cavity excavation operation. However, in cement prey operations the marketable trade good is cement, which is a man-made mineral mixture of assorted chemical components. A valid process to utilize cement market value for delegating economic value to different block chemical components is nonexistent, hence, the transition will decidedly take to an unacceptable and unrealistic analysis.

The preprocess blending of high and low class stuffs is common in unfastened cavity excavation operations. However, the end is to accomplish required content of a individual metal in the natural stuffs that are fed to one of the multiple procedures in the processing works. However, in cement fabrication operations assorted natural stuffs are fed to a individual line procedure in the cement kiln. As such, natural stuff blending to accomplish the needed per centum content of more than one chemical component is the Southern Cross of a cement fabrication operation. Therefore, delegating a individual economic value to a given block will supply inconsistent and uncomplete input to the intermixing optimisation theoretical account.

In cement prey operations, even if a block is low in CaO, it could be high plenty in SiO2, Fe2O3, or Al2O3, therefore it becomes a campaigner for excavation, blending and processing in the works. As such, a differentiation of an ore or waste block is non applicable because complete use of the limestone militias through blending is practiced in the industry.

Therefore, it may be concluded that the execution of the production programs developed utilizing economic value block theoretical accounts will be impossible due to unacceptableness of the industry and this disparity establishes the demand to develop a tool turn toing the demands of cement fabrication operations ( Rehman, 2010 ) .

Keeping in position the defects of the earlier surveies, in this research a bi-objective modified non-dominated screening familial algorithm attack is developed to seek for perchance optimal long term production agenda that minimizes the cost of providing quarry natural stuffs and the divergence from the needed natural stuffs quality parametric quantities. These long term production agendas are used as an input to an MILP preparation for bring forthing optimum and operationally applicable short term production agendas ensuing in a consonant strategic and tactical optimum production programming of cement prey operations.

Proposed Research

Problem Definition

A production agenda is constructed by using production restraints to quarry sequencing. Developing a solution to the production programming job is critical to the being of a profitable operation. Optimization of a mine agenda over a period is must if one is to recognize maximal net income from the excavation operation. Mine contrivers frequently seek to optimise production programming and sequencing with regard to a given standard. The procedure involves sequencing of ore blocks to be mined in each period over the life of the undertaking topic to precedency, production capacity, and intermixing restraints.

Production programming of natural stuffs in cement prey operations is chiefly concerned with developing a depletion sequence, get downing from the initial status of the sedimentation to the concluding prey bounds. However, there is an built-in undertaking of readying of natural mix from the run-of-mine stuff before cement production. Normally, a few stone units in a prey contain suited components to run entirely. Therefore, the normal procedure requires intermixing of different natural stuffs and, if required, with the disciplinary stuff or additives from the market. Detailss associating natural mix design are presented in Appendix 1. Objective of the production programming is to minimise the cost of providing natural stuffs, while fulfilling natural stuffs measure and quality demands.

Optimum long term production programming of prey operations for the full life of the sedimentation can be classified as a combinative hunt for the best sequence of excavation. Therefore, the development of an optimum long term production program has been a challenge due to the being of 1000s of blocks represented as whole number ( 0-1 ) variables in the mathematical scheduling preparation. Procedures like block collection, Lagrangian relaxation, and sequencing algorithms have been used to cut down whole number variables and undertake the block sequencing and precedency restraints to bring forth solutions within sensible clip ( Ramazan, 2004 ; Caccetta, 2003 ; Johnson, 2002 ; Asad, 2001 ; Dagdelen, 1987 ) .

However, application of such processs may take to a via media on the optimum solution. An alternate to these processs is using the evolutionary attacks like familial algorithms to develop optimum long term production.

Ideally, upon accomplishing an optimum long term production agenda that identifies the period-wise ( normally per twelvemonth ) blocks mining sequence fulfilling procedure measure and quality demands, the following challenge is to synchronise the monthly/weekly/daily production from identified blocks such that overall operational aims are accomplished ( Smith, 1998 ) . This leads to the following two statements:

Short term production agenda shall see the micro-level operational restraints such as equipment handiness ( mining capacity ) , material transit profiles, natural mix reserve capacity, additives handiness, etc. for the development of an optimum block excavation sequence as a subset of the annual block limits projected in the long term program.

Short term production agenda considers fewer blocks or whole number variables as compared to the long term production agenda, therefore accomplishing an optimum solution in sensible clip is possible. However, the theoretical account shall include the harsh blocks precedency restraints.

Therefore, a production agenda must take into consideration both long term planning and short term operational undertakings. Computerized systems consisting a figure of specially prepared plans to work out prey planning and production programming jobs are available for illustration FLS-CombiLogic production proctor ( Anonymous, 1988 ) , CADE and QSO systems and optimal blend scheduling algorithm ( Asad, 2001 ) etc.

Restriction of these systems and implicit in algorithm is that they are designed either for long or short term prey planning. A synchronism of strategic long term programs with operational short term production program has non been done before for cement prey operations. Therefore, there is a demand to develop quarry production planning theoretical accounts turn toing this synchronism of the optimum long and short term production programming.

Research Aims

This proposed research aims at developing theoretical accounts for optimal long and short term production programming of cement prey operations. The theoretical accounts honor the aims of both long and short term operational control programs as discussed earlier. It is expected that by making a coherence between long and short term production programs, the likeliness of successful execution, acceptableness, and attainment of the promised value of the optimum programs will increase. The theoretical accounts attempt to accomplish the undermentioned salient aims:

To minimise the cost of natural stuff from the prey and the usage of additives.

To maximise the life of prey.

To fulfill the natural stuff measure ( production capacity ) demands.

To fulfill the natural stuff quality demands.

To fulfill the prey incline and sequencing demands.

Research Methodology

The research aims are accomplished in two stairss:

The development of optimum multi-period long term production program utilizing a bi-objective familial algorithm. The algorithm minimizes the cost of providing natural stuffs from a individual beginning i.e. limestone prey, hence, it besides attempts to minimise the cumulative one-year divergence of major oxides including SiO2, CaO, Al2O3, and Fe2O3 from the specified quality demands. The familial algorithm based long term production planning theoretical account besides satisfies the prey production capacity restraints.

The development of optimum multi-period short term production program utilizing a assorted whole number additive scheduling ( MILP ) theoretical account. While taking the optimum long term production program as in input, the aim of the MILP based theoretical account is to develop a natural mix reserve by minimising the cost of providing natural stuffs from the two beginnings, including, limestone prey and additives purchased from the market.

The methodological analysis for developing optimum production agendas for cement prey operations is alone, because:

Not merely, it synchronizes the long and short term production agendas, but besides, it combines the pertinence of bi-objective familial algorithm i.e. an evolutionary attack with MILP based mathematical preparation i.e. conventional optimisation attack.

While using the familial algorithms for developing long term program, the attack evades the consideration of 1000s of prey blocks as double star ( 0/1 ) variables, and so generates more meaningful and acceptable short term production program utilizing MILP theoretical account, which considers comparatively fewer prey blocks ( merely identified in the long term program for a peculiar twelvemonth ) as double star ( 0/1 ) variables, therefore promises the solution within sensible clip.

Significance of Research

Cement is a cardinal building stuff and one of the of import mineral trade goods produced worldwide. Worldwide cement production in the twelvemonth 2009 was about 2.6 billion dozenss ( U.S.G.S, 2009 ) and it is dispersed really unevenly among more than 150 states. Figure 1.3 nowadayss world-wide cinder production capacities in the twelvemonth 2009 ( U.S.G.S. , 2009 ) . Current annual end product of cement is sufficient to do about 2.5 tons/year of concrete for every individual on the planet ( Oss, 2005 ) .

Pakistan is amongst top 20 ( 20 ) cement bring forthing states. Annual production capacity of cement in Pakistan was about 16.38 million dozenss in the twelvemonth 2000. The installed production capacity has expanded to 44.71 million dozenss per annum by the twelvemonth 2009-2010 ( APCMA, 2010 ) . Figure 1.4 nowadayss one-year production capacity in Pakistan. On norm, 1.6 dozenss of natural stuff is required to bring forth one ton of cement, hence, about 71.54 million dozenss of natural stuffs is required yearly to run into the installed production capacity.

Cement industry has witnessed tremendous betterments in treating engineering to bring forth low cost merchandise. Presently, it is concentrating on minimisation of the cost of natural stuffs production through optimal production programming. It has realized the impact of optimum prey production agendas guaranting supply of natural stuffs, which is reflecting a possible economy of Rs. 71.54 million per twelvemonth, by the cement industry in Pakistan entirely ; if a cost is simply reduced by Rs. 1/ton of natural stuffs. Optimum production programming besides ensures a coherence between planning and operational undertakings, therefore, a better technology control and a smooth cement fabrication procedure.

Thesis Organization

A brief debut of assorted optimisation techniques and a elaborate description of optimisation techniques used in this thesis, i.e. familial algorithms and whole number scheduling, are presented in chapter 2. Detailss of a bi-objective familial algorithm developed for long term production scheduling are presented in chapter 3. An illustration of working of bi-objective familial algorithm utilizing sample informations set from a block theoretical account of a limestone sedimentation in Pakistan is presented in chapter 4. Detailed description of a assorted whole number additive programming theoretical account developed for short term production programming is presented in chapter 5. The bi-objective familial algorithm and assorted whole number additive scheduling theoretical account is applied on a existent life informations set of DG Khan cement company prey operations ( located in Punjab, Pakistan ) and the consequences are given in chapter 6 followed by a sum-up of the work and a treatment on farther research waies in the concluding chapter 7.