OR-Exchange - latest questionshttp://www.or-exchange.com/feeds/rss?Ask and answer questions.enTue, 15 Apr 2014 05:43:12 -0400Automated Trading Strategies really works?http://www.or-exchange.com/questions/9497/automated-trading-strategies-really-works<p>I'm a OR student and I would like to invest some time to learn "automated trading strategies" based on OR algorithms.</p>
<p>My professor proposed me some thesis about this, and, according with its backtesting, it seems perform very well.
So, this convinced me to invest some time learning more about this.
Now, my question is: All this thing is just a "gold rush" or, in the real world there are really some hedge funds that use this?
I know, that in this field everything is uncertain, but there are someone who really win? Could you provide me something that prove it?
Does it worth it investing time on it?</p>
<p>Edit:
2) Someone knows Brokers who allow API access? I have found only "Interactive Broker" but they require $100k to open an account </p>x3d0Mon, 14 Apr 2014 13:55:44 -0400http://www.or-exchange.com/questions/9497/automated-trading-strategies-really-worksoperations-researchtradingfinanceSolved with 0 MIP simplex iterations and 0 branch&bound nodes, so what did the solver do?http://www.or-exchange.com/questions/9502/solved-with-0-mip-simplex-iterations-and-0-branchbound-nodes-so-what-did-the-solver-do<p>Hi,</p>
<p>I have an algorithm that decompose a MILP model in lots of smaller integer problems. I expected these problems to be easy to solve and, in fact, for most of them, the solver quicly finds the optimal solution stating that it used "0 MIP simplex iterations" and generated "0 branc&bound nodes".</p>
<p>The solver I've been using is CPLEX.</p>
<p>My question is: if the solver does not employ either the simplex or the B&B scheme, then how does it actually solve the problems?
I think it just uses some presolving heuristics, or it solves them by inspection.
Would I be correct?</p>
<p>TIA</p>Andrea TTue, 15 Apr 2014 05:43:12 -0400http://www.or-exchange.com/questions/9502/solved-with-0-mip-simplex-iterations-and-0-branchbound-nodes-so-what-did-the-solver-docplexmixed-integer-programmingWhere Can I Find Polished Analytics Case Studies on "The Americas"?http://www.or-exchange.com/questions/9494/where-can-i-find-polished-analytics-case-studies-on-the-americas<p>Hi: I'm looking for polished (i.e. HBR-style) case studies that deal with at least one (ideally all three) of the main phases of Analytics: description, prediction, and prescription. That is, I want a case study that will require students to use stats for analysis, forecasting, etc., or maybe some data-mining (e.g. clustering, classification), or some optimization, or a combo of these. The main requirement is that the story behind the case be about a company in the Americas. A global/cross-country business situation would be even better. After looking through the list of HBR cases, I haven't found anything I like. I'll now move on to the other usual suspects (e.g. Darden), but if anyone has a tip on what publishers have a bigger selection of O.R.-ish cases, I'd greatly appreciate it. Thanks in advance!</p>Tallys YunesFri, 11 Apr 2014 17:13:14 -0400http://www.or-exchange.com/questions/9494/where-can-i-find-polished-analytics-case-studies-on-the-americasanalyticsglobal-businesslatin-americacase-studyComposite piecewise functions?http://www.or-exchange.com/questions/9489/composite-piecewise-functions<p>Hi all,</p>
<p>I have a simple, but tricky problem.
Please let me know if anyone knows the answer of below problem.</p>
<p>V(t) = 0, if 0<= t < 300,
-1000, if t >= 300.</p>
<p>g(t) = -0.25t + 60, if 0 <= t < 100,
35, if t >= 100.</p>
<p>Then, what is V(t + g(t)) = ?
(where t is a time variable, so t >= 0)</p>
<p>I don't know the exact way of composition of two piecewise functions in this case.
(ranges are different)</p>
<p>Thank you.</p>GitaeWed, 09 Apr 2014 22:12:26 -0400http://www.or-exchange.com/questions/9489/composite-piecewise-functionsfunctionHow relevant are stochastic modeling and stochastic control to data science?http://www.or-exchange.com/questions/9471/how-relevant-are-stochastic-modeling-and-stochastic-control-to-data-science<p>Hi,</p>
<p>I am a graduate student at Stanford, and I was wondering how much knowledge of Stochastic Modeling and Stochastic Control would help me as I get into data sciences as a career. Have to decide on my coursework for the quarter. Would appreciate if somebody could shed some light on how widely used these are.</p>
<p>Thanks!</p>shadowblade360Fri, 04 Apr 2014 20:02:45 -0400http://www.or-exchange.com/questions/9471/how-relevant-are-stochastic-modeling-and-stochastic-control-to-data-sciencestochastic-optimizationmachine-learningdata-mining465 Too much data on this linehttp://www.or-exchange.com/questions/9479/465-too-much-data-on-this-line<p>Dear Sir, </p>
<p>I am learning how to use the GAMS program and I am meeting a problem about the setting of the dataset. The situation is as following:</p>
<p>The original dataset I used is </p>
<pre><code>/*Import the data table*/
table act(obs, j) input output table
INCLUDE D:\testshare.csv
,x11,x12,x13,x21,x22,x23,x31,x32,x33,y11,y12,gy21,gy22,by2,gy31,gy32,by3,x
s1,xs2,en1,en2
f1,698,356,568,235,178,568,254,48,475,145,152,135,165,89,325,154,1398,3654
,1952,0.89,0.95
</code></pre>
<p>After I add a dimension into the dataset without changing anything else, an error information shown as </p>
<pre><code>/*Import the data table*/
table act(obs, t, j) input output table
INCLUDE D:\testsharemalm.csv
,x11,x12,x13,x21,x22,x23,x31,x32,x33,y11,y12,gy21,gy22,by2,gy31,gy32,by3,x
s1,xs2,en1,en2
f1.2001,698,356,568,235,178,568,254,48,475,145,152,135,165,89,325,154,1398
,3654,1952,0.89,0.95
**** $465
</code></pre>
<p>and the error information for $465 is "465 Too much data on this line". I don't know what happened to my setting of data. Would you please tell me how to solve this problem? Thanks a lot.</p>
<p>Your sincerely,</p>
<p>Chih Cheng Chen</p>iebcgcMon, 07 Apr 2014 06:16:25 -0400http://www.or-exchange.com/questions/9479/465-too-much-data-on-this-linemuchdatatooseetingWhich (heuristic) approaches to multi-stage optimization problems do you use?http://www.or-exchange.com/questions/9439/which-heuristic-approaches-to-multi-stage-optimization-problems-do-you-use<p>Multi-stage optimization, or optimization over time, or ..., in the sense I would like to discuss is to make decisions for a planning horizon, where decisions at a stage have immediate consequences only to the state in the next stage. A simple example is lot-sizing, where you take decisons about how much to produce and store at every stage; what you store today will be available tomorrow. The equation linking "today and tomorrow" is the inventory balance, which does not link today and the day-after-tomorrow. I am particularly interested in integer programming formulations (the matrices of these formulations will have something like a "staircase form") and/or heuristics you use to solve such problems (rolling horizon etc.). Also interested in pointers to the literature. Thanks! </p>
<blockquote>
<p><strong>Edit</strong>: We are testing, among others, primal heuristics for MIP formulations of such problems, so I would be also interested in LP files. </p>
</blockquote>Marco LuebbeckeWed, 26 Mar 2014 20:18:33 -0400http://www.or-exchange.com/questions/9439/which-heuristic-approaches-to-multi-stage-optimization-problems-do-you-useheuristicsinteger-programmingmatrixmulti-stage-decisionsany users of the Standard Point Location Code (SPLC) database?http://www.or-exchange.com/questions/3733/any-users-of-the-standard-point-location-code-splc-database<p>The National Motor Freight Traffic Association publishes a database of Standard Point Location Codes (SPLC). They say:</p>
<blockquote><i>The SPLC is designed to provide each point originating freight and each point receiving freight in North America with a unique code number so constructed as to identify the point with its geographic location, using two digits to identify State, County and City and three digits to identify Sub-Code.</i></blockquote>
<p>Sub-codes can be used to identify things I'm interested in, such as Plant sites, Warehouses, Docks, and Piers, as well as things I'm not so interested in such as Resorts, Parks, Race Tracks, Amusement Centers, and Zoos. (Well, I might be interested in some of those, just not while I'm at work.)</p>
<p>Product's described here:
<a href="http://www.nmfta.org/Pages/Splc.aspx">http://www.nmfta.org/Pages/Splc.aspx</a></p>
<p>I'm wondering if any of you in transportation and logistics have used this database, and have found there to be substantive data at the sub-code level. Essentially, I'm looking for a Point of Interest database that's highly relevant for motor freight: production facilities, distribution centers, warehouses, etc.</p>
<p>I'd be happy to learn of your experiences with the SPLC, or hear about any similar products or efforts to compile such data.</p>
<p>Thanks, Eric</p>eric theiseTue, 20 Sep 2011 15:51:35 -0400http://www.or-exchange.com/questions/3733/any-users-of-the-standard-point-location-code-splc-databaselogisticstransportationlocationhow to create such constraint in OPL of cplexhttp://www.or-exchange.com/questions/9392/how-to-create-such-constraint-in-opl-of-cplex<p>Hi ,all</p>
<p>I want to create a constraint in OPL.I don't know how to ?</p>
<pre><code>tuple JUP
{
string p;
string u;
} ;
{JUP} jup=...;
dvar interval TT[o in jup];
</code></pre>
<p>I want to create a constraint </p>
<pre><code>forall(o in jup:o.p==p1)
the number of 'startOf(TT[o])<=5' will be less than 10
</code></pre>
<p>Could you help me?</p>
<p>Thanks all.</p>spzcwgWed, 12 Mar 2014 21:45:20 -0400http://www.or-exchange.com/questions/9392/how-to-create-such-constraint-in-opl-of-cplexoplcplexconstraintunimodular vs NP hardhttp://www.or-exchange.com/questions/9449/unimodular-vs-np-hard<p>Dear All,
I have a MIP (mixed interger programming model) network interdiction problem ( Ref: Deterministic Network Interdiction, K.R. Wood). It is proved to be NP hard. In the model there are contraints that come from the maxflow problem. I have implemted the model and when I solve the LP relxation of the problem I get a solution same as that of the MIP.
``I am stumped. I am confused as to where to start looking. I looked at my implementation and it looks correct. So my question is can the unimodularity property of the max flow problem in the orignial MIP play a role? And it it true that a NP hard problem cannot be unimodular?</p>
<p>Please help me here, I have been thinking about this for more than two weeks now.</p>
<p>Than you in advance</p>satishSun, 30 Mar 2014 12:23:33 -0400http://www.or-exchange.com/questions/9449/unimodular-vs-np-hardmiplpunimodularrelaxationif-then-and-or constrainthttp://www.or-exchange.com/questions/9443/if-then-and-or-constraint<p>i have a conditional constrain. this constraint must change to linear constraint.
please help me.</p>
<p>if x_1> x_2> x_3 then y_1> y_2> y_3
and
if x_1<=x_2<=x_3 then y_1<= y_2<= y_3</p>
<p>x_1,x_2,x_3,y_1,y_2,y_3 is a variable</p>
<p>thank you</p>eric kristantoSat, 29 Mar 2014 08:41:55 -0400http://www.or-exchange.com/questions/9443/if-then-and-or-constraintconditionalconstraintLinearizing product of continous and binary using BigMhttp://www.or-exchange.com/questions/9453/linearizing-product-of-continous-and-binary-using-bigm<p>Dear All,
This question is a contunation of my previous question <a href="https://www.or-exchange.org/questions/9449/unimodular-vs-np-hard">link text</a>.
I have a multiplication of a binary variable and a continous variable x and A respectively. I have linearized it by introducing another continous varibale z=x*A. I had the follwing constraints:</p>
<ol>
<li>z <= M * x </li>
<li>z <= A </li>
<li>z >= A - (1 - x)M</li>
</ol>
<p>Where M is BigM a big number. Now when I set M=1 and relax the binary x variable I get LP relaxation = MIP solution. But, as I increase M I get LP relaxation < MIP which is desired for a minimization problem.</p>
<p>I really need your help on this. Thannk you</p>satishSun, 30 Mar 2014 15:21:36 -0400http://www.or-exchange.com/questions/9453/linearizing-product-of-continous-and-binary-using-bigmmiplinearization-techniquerelaxationWhat is the size of the search space of a vehicle routing problem?http://www.or-exchange.com/questions/9420/what-is-the-size-of-the-search-space-of-a-vehicle-routing-problem<p>Presume we have a normal Vehicle Routing Problem (VRP) with <code>n</code> customers (= locations) and <code>v</code> vehicles. Each vehicle has 1 route. <em>In how many different ways can we assign those <code>n</code> customers to the <code>v</code> routes? What's the formula?</em></p>
<p>For example. If we have 2 customers (A, B) and 2 vehicles (1, 2), this is a list of all possible combinations:</p>
<ol>
<li>1AB - 2</li>
<li>1BA - 2</li>
<li>1A - 2B</li>
<li>1B - 2A</li>
<li>1 - 2AB</li>
<li>1 - 2BA</li>
</ol>
<p>So the formula with <code>n=2</code> and <code>v=2</code> should return <code>6</code>.</p>Geoffrey De SmetTue, 25 Mar 2014 08:08:20 -0400http://www.or-exchange.com/questions/9420/what-is-the-size-of-the-search-space-of-a-vehicle-routing-problemvehicle-routinghow to use relax a constraint in CP optimizer?http://www.or-exchange.com/questions/9405/how-to-use-relax-a-constraint-in-cp-optimizer<p>Hello all:</p>
<p>I want to use relax for a constraint in CP optimizer,so I labeled this constraint.
But it does not work.</p>
<pre><code>using CP;
forall(jp in book)
{
DueC:endOf(JobProcess[jp])<=2;
}
</code></pre>
<p>How to releaze this funciton? Thanks. <br>
</p>spzcwgMon, 17 Mar 2014 22:06:18 -0400http://www.or-exchange.com/questions/9405/how-to-use-relax-a-constraint-in-cp-optimizercplexSearch Algorithmhttp://www.or-exchange.com/questions/9399/search-algorithm<p>I am working on a research problem, which requires a search algorithm over integer.
I have a code to calculate a cost with given \(x\). \(x\) can be only a positive integer.
I can calculate \(f(x)\), numerically if \(x\) is given. (calculus doesn't work) </p>
<p>Fortunately, \(f(x)\) is convex in \(x\). </p>
<p>But due to the fact that the bigger \(x\) requires the larger matrix to deal with, and consequently, it requires much longer computation time. </p>
<p>We don't have upper nor lower bound. </p>
<p>How can I find \(x\) that minimize the cost function \( f(x)\), effectively ?</p>
<p>I found golden section search and fibonacci search algorithms.
But, they assume that there are upper and lower limits so that the optimal point is located in-between.
And, they are not for integer points.
Since large \(x\) leads a long computation time, I don't want to set a huge number as my upper limit. </p>
<p>I developed my own search algorithm, but I am not satisfied with my algorithm.<br>
</p>ksphilSun, 16 Mar 2014 21:34:00 -0400http://www.or-exchange.com/questions/9399/search-algorithmalgorithmssearching[ANN] The conic benchmark libraryhttp://www.or-exchange.com/questions/9409/ann-the-conic-benchmark-library<p>I would like to announce the <strong>conic benchmark library</strong> which is a library of conic optimization test instances. Currently it emphases conic quadratic (aka. second order cone) problems possibly including integer constrained variables. It is available at the <a href="http://cblib.zib.de/">Zib CBLIB site</a>.</p>
<p>We hope to add some semidefinite optimization problems in the near future. Finally, I strongly urge you to contribute more instances so it can become even more relevant. Contributing interesting instances is the best way to influence the development of the conic optimization solvers in a direction you want!</p>Erling_MOSEKThu, 20 Mar 2014 02:23:39 -0400http://www.or-exchange.com/questions/9409/ann-the-conic-benchmark-libraryconicquadraticbenchmarksocpSyntax for i belongs to set I, but not equal to m (in OPL)http://www.or-exchange.com/questions/8987/syntax-for-i-belongs-to-set-i-but-not-equal-to-m-in-opl<p>Hello all,</p>
<p>I am writing a code for a formulation in CPLEX. I want to write a constraint which involves i belongs to set I, but not equal to m, where m is a number from the subset of I.</p>
<p>The ampl syntex for the above one is
subject to buyer{i in m}: sum{i in I: i<>m}....</p>
<p>Can anyone help me in CPLEX syntex for the same..</p>
<p>Thanks. </p>RatnamFri, 03 Jan 2014 05:02:47 -0500http://www.or-exchange.com/questions/8987/syntax-for-i-belongs-to-set-i-but-not-equal-to-m-in-oploplsyntaxACO on SCPhttp://www.or-exchange.com/questions/9401/aco-on-scp<p>I am an undergrad student and I am currently working on my Special Problem entitled "Application of Ant Colony Optimization Metaheuristic on Set Covering Problems."</p>
<p>My problem is that I need a program that can solve SCP using ACO algorithm. There are lots of ACO program there but most of them are solving Travelling Salesman Problem. I also found a useful article about new ideas for applying ant colony optimization to the set covering problem, it has the codes that I will be needing but my problem is I can't decode it in any other program. I just don't know how. </p>
<p>So, can someone help me by giving a program about ACO solving SCP? It would be a really great help.</p>JersssMon, 17 Mar 2014 01:07:43 -0400http://www.or-exchange.com/questions/9401/aco-on-scpprogrammingmetaheuristicoptimizationacoscpR, Matlab or Python for solving optimization problemhttp://www.or-exchange.com/questions/9345/r-matlab-or-python-for-solving-optimization-problem<p>To solve the small-to-mid scale optimization problem, which software is better, R, MATLAB or Python? If MATLAB is too expensive, will R and Python be used as a reliable alternative option?</p>winecodingMon, 03 Mar 2014 08:59:20 -0500http://www.or-exchange.com/questions/9345/r-matlab-or-python-for-solving-optimization-problempythoninteger-programmingoptimization-softwareoptimizationsoftwareRemote sensing imagehttp://www.or-exchange.com/questions/9394/remote-sensing-image<p>Hi there,
I am always confusing by the words of 'pattern' and 'texture' in image classification. How can I understand these? I am happy if anyone answer my question.</p>HaniThu, 13 Mar 2014 11:20:26 -0400http://www.or-exchange.com/questions/9394/remote-sensing-imagelandcoverclassificationBenchmark instances for Vehicle Routing Problemhttp://www.or-exchange.com/questions/9388/benchmark-instances-for-vehicle-routing-problem<p>I am running some benchmarks on Vehicle Routing Problem with Time Windows, using the (Gehring and Homberger 1999) instances.
But these problems have only up to 1000 customers.</p>
<p>Does anyone know of a data set for VRPTW that goes up to 10'000 or at least higher than a 1000 ?</p>digyTue, 11 Mar 2014 13:48:04 -0400http://www.or-exchange.com/questions/9388/benchmark-instances-for-vehicle-routing-problemvrpvrptwGeneration Expansion Problem (GEP) with multiple objectiveshttp://www.or-exchange.com/questions/9308/generation-expansion-problem-gep-with-multiple-objectives<p>Dear friends,</p>
<p>I'm looking for a sample codes (preferably GAMS) on Electricity Generation Expansion Problem
(GEP) with multiple objectives. Another resource that I am looking for is
the GEP problem under demand growth uncertainty</p>
<p>Any suggestion for relevant resources or projects is highly appreciated.</p>
<p>Thanks in advance for your kind support.</p>
<p>Regards,</p>
<p>Jozdalga</p>JozdalgaFri, 21 Feb 2014 15:51:08 -0500http://www.or-exchange.com/questions/9308/generation-expansion-problem-gep-with-multiple-objectivesuncertaintystochastic-optimizationmultiobjective"OR" between constraintshttp://www.or-exchange.com/questions/9383/or-between-constraints<p>In all optimization models, we have "And" between all constraints i.e all constraints should be satisfied all together. Do you know any practical optimization model in which we have "OR" between at least sum of constraints? for example:</p>
<pre><code> Min x1+x2 subject to {2x1+x2<4 OR {x1<3 AND x1,x2>0}}
</code></pre>Reza SalehizadehMon, 10 Mar 2014 16:52:04 -0400http://www.or-exchange.com/questions/9383/or-between-constraintsmathematical-modelingCP Problemhttp://www.or-exchange.com/questions/9379/cp-problem<p>I'm still a beginner in CPLEX and I have a program which I want to model with the constraints and objective function.
I have a cloud architecture which contain J data center and each data center has a set of resources, and a set of requests and each request has its requirements.
To solve this problem, i created a boolean varibale y(ijk) which take 1 if the resource k in the center j can fit my request i and 0 otherwise. So from the set of resource where the y(ijk)==1, i have to select the first site and the first resource to serve my request i. That's why i created two other boolean variable z(ij) which take 1 if the center n° j is the first center and x(ijk) is equal to 1 if k is the the first resource of the first center j that fit my request. I chose for this this two constraints:
(j + 1) zij <= (s + 1)<em> yijk
(k + 1) xijk <= (r + 1) * yijk</em> zij. But in vain i coudn't select the adequate resource. Can someone help me please. Thank's in advance.</p>OualiFri, 07 Mar 2014 05:45:15 -0500http://www.or-exchange.com/questions/9379/cp-problemcplexCPLEX Problemhttp://www.or-exchange.com/questions/9378/cplex-problem<p>Hello, I'm still a beginner in CPLEX and I have a program which I want to model with the constraints and objective function. In fact, the principe of the program is as follows:</p>
<p>"I have a topology of a cloud that contains J data centers and each center has a set of resources. Each resource has a processing capacity and bandwidth. I also have a set of requests that I want assigned to the appropriate resources. Each application has a capacity and bandwidth and must be served by the first adequate resource."</p>
<p>But I have not been able to select the first resource in this problem – and I need your help.</p>
<p>You will find attached the model I designed and thank you in advance.</p>
<p></p><center>
<img alt="alt text" src="http://image.noelshack.com/fichiers/2014/10/1394112508-modele.png">
</center><p></p>OualiThu, 06 Mar 2014 10:07:47 -0500http://www.or-exchange.com/questions/9378/cplex-problemcplexProbleme Cplexhttp://www.or-exchange.com/questions/9376/probleme-cplex<p>Bonjour, je suis encore débutante en CPLEX et j'ai un programme que je veux modéliser à l'aide des contraintes et une fonction objective.
Au fait, le principe du programme est le suivant: " J'ai une topologie d'un cloud qui contient J centres de données et dans chaque centre un ensemble de ressources. Chaque ressource dispose d'une capacité de traitement et d'une bande passante. J'ai aussi un ensemble de demandes que je veux affectées aux ressources qui conviennent. Chaque demande dispose d'une capacité et d'une bande passante et doit être servie par la première ressource adéquate."
Mais j'ai pas arrivé à résoudre ce problème. et j'ai besoin de votre aide.
Vous trouverez ci-joint le modèle que j'ai conçu et merci d'avance. <img alt="alt text" src="http://www.noelshack.com/2014-10-1394112508-modele.png"></p>OualiThu, 06 Mar 2014 09:35:54 -0500http://www.or-exchange.com/questions/9376/probleme-cplexcplexDo you want to test post here?http://www.or-exchange.com/questions/2547/do-you-want-to-test-post-here<p>We now have new software. If you would like to put a test answer here, you are welcome to. </p>Michael TrickFri, 08 Apr 2011 13:57:32 -0400http://www.or-exchange.com/questions/2547/do-you-want-to-test-post-heremetanew-or-exchangesite-administrationParallel and distributed Optimizationhttp://www.or-exchange.com/questions/9029/parallel-and-distributed-optimization<p>Folks, I have worked on building and deploying optimization models in one machine and never worked or completed coursework on parallel optimization. It seems quite interesting and would like to know more about it. Please may I request OR folks here to suggest me resources on where to start. I would like to know how to set up a parallel computing platform (given 2 or more machines readily available) and how to programming a simple LP using existing parallel computing setup.</p>
<p>Also is it possible to run parallel optimization models on Hadoop (either mapreduce or YARN)?</p>PavanWed, 08 Jan 2014 12:21:10 -0500http://www.or-exchange.com/questions/9029/parallel-and-distributed-optimizationparallel-programmingoptimizationbig-datamathematical-modelingOut of memory error with CPLEXhttp://www.or-exchange.com/questions/9359/out-of-memory-error-with-cplex<p>Hi everyone, </p>
<p>Every time I run my model (after 30 min with gap 99.8%), face "error 1001: out of memory" error!
I set memory emphasis parameter to 1, working memory parameter to 1 and node file parameter to 2, to let cplex store node files on disk. However, the physical memory usage is yet at a high level and no node file is stored!!!
Do you have any suggestions to carry out this problem? </p>
<p>for more information:
i run the model on computer with 4G RAM, the version of CPLEX that i'm using is 12 and the log file output is: </p>
<p>Presolve time = 145.91 sec.
Probing time = 0.95 sec.
Clique table members: 1271.
MIP emphasis: balance optimality and feasibility.
MIP search method: dynamic search.
Parallel mode: deterministic, using up to 4 threads.
Root relaxation solution time = 7.18 sec.</p>
<p>Cover cuts applied: 263
Implied bound cuts applied: 19854
Flow cuts applied: 1356
Mixed integer rounding cuts applied: 6973
Zero-half cuts applied: 18
Gomory fractional cuts applied: 93</p>
<p>Total (root+branch&cut) = 1642.10 sec.</p>
<p>Warning: MIP starts not constructed because of out-of-memory status.</p>
<p>Thanks</p>dinaTue, 04 Mar 2014 04:34:43 -0500http://www.or-exchange.com/questions/9359/out-of-memory-error-with-cplexerror1001cplexEffect of form of equations on IP solvinghttp://www.or-exchange.com/questions/9341/effect-of-form-of-equations-on-ip-solving<p>I've been dealing with a family of hard 0/1 IP's all of which are of the form Ax = b (no inequalities except for variable bounds). When I give the problems to CPLEX in their original form it always reports that its been able to perform a number of coefficient reductions (though I don't know which ones. That would be interesting to find out). After looking at the form of the matrix A I noticed that it was possible to do a number of row operations on A (and doing the same operations to b) which would greatly increase the sparsity of A. Now when I give the modified equations to CPLEX it never reports any coefficient reductions, and, in general solves them about 25% to 50% faster. So this brings up the question -- is there are systematic research out there which is able to describe "better" forms of equations? I just went on general idea that sparser was better, but that still seems a bit vague.</p>VictorSMillerSun, 02 Mar 2014 15:18:42 -0500http://www.or-exchange.com/questions/9341/effect-of-form-of-equations-on-ip-solvingmipcplex