全傑科技 Softhome - Evolver

Are you working on complex engineering, scheduling, or resource allocation problems? Do you have problems that Excel’s Solver claims can’t be solved? You need the most innovative solver technology on today’s preferred platform. Look no further than Evolver™, the revolutionary Windows-based problem-solving tool that works as an add-in to Microsoft Excel. Evolve the Optimal Solution! Evolver is an optimization add-in for Microsoft Excel?/sup>. Evolver uses innovative genetic algorithm (GA) technology to quickly solve complex optimization problems in finance, distribution, scheduling, resource allocation, manufacturing, budgeting, engineering, and more. Virtually any type of problem that can be modeled in Excel can be solved by Evolver, including previously unsolvable problems! Evolver requires no knowledge of programming or GA theory and ships with a fully illustrated manual and several examples. Evolver is available in three versions: Standard, Professional, and Industrial. The Professional and Industrial versions have increased problem capacities and advanced features. Evolver 4.0: The Most Powerful Evolver Ever! Anyone who uses Evolver will be amazed at the new Evolver 4.0! Evolver 4.0 is easier to use, with redesigned toolbars and dialog boxes. Larger problems than ever before can be defined. For faster and better solutions, the Professional and Industrial versions support auto-mutation rates and selectable genetic operators. These versions can adjust the mutation rate during an optimization and focus in on better solutions. They can also determine which genetic operator - heuristic crossover, extinction, boundary mutations, and more - will find the best solution to your optimization problems. Many Evolver 4.0 users report that they are getting better answers to their optimization problems that they have ever gotten before - with any optimizer. And they're getting these answers much, much faster! Genetic Algorithms and Evolution In nature, we can observe 'problems' being solved through evolution; species evolve as they adapt to dynamic environments. We extend this observation into the world of computer software via Evolver, the most powerful optimization package available. Evolver uses innovative GA technology to create environments where possible solutions continuously crossbreed, mutate, and compete with one another, until they 'evolve' into the best solution. As a result, Evolver can find optimal solutions to virtually any type of problem, from the simple to the most complex. The Right Tool For The Job Evolver includes six highly specialized algorithms for different types of problems: Recipe - a set of variables which can change independently. Like ingredients in a cooking recipe, they are combined to find the best mix. Grouping - a collection of elements to be placed into groups. For example, place workers into groups to perform different jobs, or group stocks into equally valued bundles. Order - an ordered list of elements. For example, the order in which to let students choose their classes, or the optimal order to execute several interdependent tasks. Budget - recipe algorithm, but total is kept constant. Project - order algorithm, but some elements precede others. Schedule - group algorithm, but assign elements to blocks of time while meeting certain constraints. Using Evolver: Building Your Model To use Evolver, you must first build a model of your existing problem using Microsoft Excel. If you already have a model, congratulations! You are ready to run Evolver. If you don’t have a model and aren’t sure where to begin, look through the twenty examples included with Evolver. Your model can use any functions (including look-up tables and if-then statements) and can even call VBA macros. Unlike traditional optimization tools, Evolver will not force you to make unrealistic assumptions in your model that make the problem abstract. Your model needs to meet these three criteria: The model contains variables that interact in some way to produce a result. The model recalculates properly for all possible combinations of variables. That result is an accurate representation of how good any given solution is. As long as these criteria are met, Evolver can intelligently recombine variables to find the optimal answer, whether it’s the best values, the right order, or optimum grouping. Sometimes it’s difficult to ensure that the second criteria listed above will be met while building a model. You can check this second criteria by telling Evolver to pause if an error occurs during an optimization. This is a good way to 'debug' your model. Using Evolver: Setting Up and Optimizing Installing Evolver is a snap. After running the installation program, the Evolver toolbar is added to Excel. To set up Evolver, click on the Evolver Settings button, then: Specify which spreadsheet cell contains the output to be optimized (the target cell). Enter the cells that Evolver can adjust (the variables), along with the solving method to use. Specify any constraints that must be met (such as $A$1 <= 5000, $A$2 * $A$3 >= 2000, etc.). You can also specify the conditions that need to be met for Evolver to stop – stop after n trials, stop after n minutes, stop when change in last n trials is less than a particular value, or stop when a particular Excel formula is true. That’s all there is to it! Click on the 'Optimize' button and Evolver will find the optimal solution to your problem! At any time you can Pause or Stop Evolver’s progress using the Evolver toolbar. Evolver can solve the most complex problems that other solvers can't! How Evolver Works Evolver solves optimization problems using genetic algorithms (GAs). In a GA, each guess or possible solution to a given problem becomes an independent 'organism' that can 'breed' with other organisms. The spreadsheet model acts as an environment for the organisms, determining which are 'fit' enough to survive based on their results. Here is a brief description of the process: Randomly generate many organisms (possible solutions), and calculate the result each organism produces. This entire 'population' of organisms is ranked from best to worst. Select good organisms and swap their variables (genes) using crossover and mutation to produce 'offspring'. If offspring do not produce a good result, two more parents are selected. If the offspring organism is good, it is re-inserted into the population. As Evolver repeats steps 2 and 3, the population 'evolves' increasingly optimal solutions. Local vs Global Solutions The methods most optimization programs use – including Solver – are 'hill climbing' methods. Hill climbing methods will start with an initial guess and proceed to find the nearest maximum (or minimum) solution. You can imagine starting at a point along a curve and moving to the left or right until you reach a peak or trough. Hill climbing will always find the best answer if (a) the function being explored is smooth, and (b) the initial variable values place you near the optimal solution. If either condition is not met, hill climbing can end up at a local solution rather than a global solution. Evolver does not use hill climbing. It uses GA technology, a stochastic directed searching technique. This method does not get stuck at local solutions, but instead looks at the entire range of possible solutions. This not only allows you to use functions such as look-up tables and if-then statements in your model that make the function discontinuous, but allows Evolver to zero in on a global solution. If you have a complex or highly non-linear problem (as many real-world problems are), you should, and often must, use Evolver. What Evolver Can Do Evolver can find optimal solutions to any function that results in a number, whether calculated in a single cell, a full sheet of equations, or the result of some large, complex macro. Evolver’s GA doesn’t need to know the specific details of your problem, or even whether it is solving linear, nonlinear, stochastic (random), combinatorial, noisy, or probabilistic functions. It doesn’t even need a starting guess to get going. Just click 'Optimize' and the rest is automatic! The Evolver Watcher During an optimization, you can bring up the Evolver Watcher™. The Evolver Watcher is a separate application that monitors your computer for all Evolver activity, whether it is being called from Excel or from your own application (see Evolver Developer’s Kit below). It allows you to change the algorithm parameters, explore the population, and display graphs, charts, and complete reports on the status of your problem. These graphs update automatically while Evolver is running. Three Versions to Choose From! Evolver is available in three versions: Standard, Professional, and Industrial. Each version has different capabilities: . Variables Auto-Mutation Rates Evolver Standard 80 No Evolver Professional 250 Yes Evolver Industrial Unlimited Yes Auto-Mutation Rates Professional and Industrial versions of Evolver feature auto-mutation rate adjustment. The mutation rate determines the amount of randomness Evolver uses when generating new solutions. Evolver Professional and Industrial can monitor progress toward an optimal solution and adjust the mutation rate appropriately. Selectable Genetic Operators Professional and Industrial versions of Evolver also allow users to select from a variety of available genetic operators, including heuristic crossover, extinction, boundary mutations, local search, and others. Choose the best operator for your specific problem to find the optimal answer more efficiently. Using a powerful auto-operator feature, Evolver can test all available operators and identify the best-performing one for a specific model. Evolver 4.0 Features True 32-bit processing for Excel 7 (Office 95) and Excel 8 (Office 97) – increases performance and allows Evolver to handle larger models more efficiently 20x average speedup over previous versions; some models run up to 100x faster! Improved interface – redesigned toolbar, dialog boxes Support for disjoint ranges and multiple sheets Pause on error and update display options Controllable via VBA macros for within Excel; can run VBA macros each recalculation (Professional and Industrial versions only) Evolver Watcher – monitors Evolver activity in Excel and custom applications; generates graphs and reports Auto-mutation rate adjustment – provides faster and better solutions (Professional and Industrial versions) Selectable genetic operators – provides more accurate solutions to various problems; auto-operator feature chooses the optimal operator (Professional and Industrial versions) Evolver Developer's Kit Also Available The Evolver Developer’s Kit (EDK) allows developers to add the power of Evolver to their custom Windows applications. The EDK can be used by any Windows programming environment that can access Dynamic Link Libraries (DLLs), such as Visual Basic, C++, Delphi and more. The EDK gives programmers access to Evolver’s powerful GA technology and allows them to use Evolver where they need it instead of linking to Excel. Programs developed with the EDK can run faster than similar models designed in Excel. EDK users can also develop new solving methods that can be accessed by Evolver. The EDK comes with complete documentation and examples in Visual Basic and Visual C++. Contact your Palisade Sales Representative for more information on the EDK and on distributing EDK-developed programs. Evolver System Requirements Minimum Platform: IBM PC compatible Pentium-equivalent or higher, 16MB RAM, Windows 95/98, NT 4.0, Windows 2000, Windows XP Recommended: 32 MB RAM, or greater Spreadsheet: Windows Excel 97, Excel 2000, Excel XP Version: 4.0 Technical Support: 30 days free. Further support available through Maintenance Plan. Demo: Web download and free demo CD with trial version available. Training: Available through Palisade's Software Training Courses.