https://training-course-material.com/index.php?action=history&feed=atom&title=Simulation_OptimizationSimulation Optimization - Revision history2026-04-09T12:41:17ZRevision history for this page on the wikiMediaWiki 1.45.1https://training-course-material.com/index.php?title=Simulation_Optimization&diff=8079&oldid=prevBernard Szlachta: /* Heuristic */2012-11-23T23:08:04Z<p><span class="autocomment">Heuristic</span></p>
<p><b>New page</b></p><div>{{Cat|OCEB T200.6 - Monitoring and Managing}}<br />
<br />
Source:http://bptrends.com/publicationfiles/01-05%20WP%20Simulation%20Optimization%20-%20April%20et%20al.pdf<br />
<br />
== Overview ==<br />
* BPMS offers simulation capabilities <br />
* Simulation is "a means to evaluate the impact of process changes and new processes in a model environment through the creation of “what-if” scenarios"<br />
* decisions can be tested before going life<br />
* simulation allow forecast with a level of uncertainty <br />
<br />
== Need for Simulation Optimization ==<br />
* analyst wants to find a set of model optimal performance (parameters and structure)<br />
* range of parameter values and the number of parameter combinations is too large for analysts to simulate all possible scenarios<br />
* they need a way to find a good solutions<br />
<br />
== Mathematical methods vs Simulation ==<br />
* many problems are too complex to be modelled in analytical way<br />
* pure optimization models alone are incapable of capturing all the complexities and dynamics of the system<br />
* simulation cannot easily find the best solutions<br />
* '''Simulation Optimization''' combines both methods (analytical optimization of the simulation or vice versa)<br />
<br />
== Heuristic ==<br />
* Greek: "Εὑρίσκω": find, discover<br />
* technique designed for solving a problem more quickly when classic methods are too slow<br />
* finding an approximate solution when classic methods fail to find any exact solution (by trading optimality, completeness, accuracy, and/or precision for speed)<br />
* e.g. rule of thumb, an educated guess, an intuitive judgement<br />
<br />
== Metaheuristic ==<br />
* designates a computational method that optimizes a problem by iteratively trying to improve a candidate solution with regard to a given measure of quality<br />
* make few or no assumptions about the problem being optimized and can search very large spaces of candidate solutions<br />
* do not guarantee an optimal solution is ever found<br />
* implement some form of stochastic optimization.<br />
(Wikipedia)<br />
<br />
== Metaheuristics ==<br />
[[File:Metaheuristics classification.svg]]<br />
<br />
<br />
== Problems with Simulations ==<br />
* Optimization models were thought to over-simplify the real problem<br />
* this was improved by research in '''metaheuristics''' along with improved statistical methods of analysis<br />
<br />
== Metaheuristics ==<br />
* Coined in 1986 by Dr. Fred Glover<br />
* "describe a master strategy that guides and modifies other heuristics to produce solutions beyond those that are normally generated in a quest for local optimality"<br />
<br />
<br />
* There are algorithms to guide a series of simulations towards good results in the absence of tractable mathematical structures<br />
* Quality of different solutions can compared<br />
* commercial products use discrete-event or Monte Carlo simulation to performs search for optimal values of input parameters<br />
* tool for commercial simulation software, employs metaheuristics ('''scatter search''', '''tabu search''', '''neural networks''')<br />
<br />
== Optimization of Simulation Models ==<br />
# develop simulation model for a system or a process<br />
# set performance measure (for possible set of choices)<br />
# find a configuration that produce good results<br />
<br />
;Extreme Methods:<br />
# trial-and-error<br />
# enumeration of all possible configurations<br />
<br />
== Applications ==<br />
* configuration of machines for production scheduling<br />
* layouts, links, and capacities for network design<br />
* investment portfolio for financial planning<br />
* utilization of employees for workforce planning<br />
* course scheduling<br />
<br />
== Black-box Model ==<br />
; Input<br />
* input parameters and/or structural design that lead to optimal performance ('''factors/levels''', '''decision variables''')<br />
; Output<br />
* performance measures ('''responses''' - used to model an objective function and constraints)<br />
; Goal<br />
* find out which factors have the greatest effect on a response<br />
* combination of factor levels that minimizes or maximizes a response<br />
* subject to constraints imposed on factors and/or responses<br />
; Constraints<br />
* constraint for both: decision variables and responses need to be formulated<br />
<br />
;Example<br />
: manufacturing facility <br />
: factors - number of machines of each type, machine settings, layout, and the number of workers<br />
: responses - cycle time, work-in-progress, and resource utilization<br />
: goal - reduce cost, minimize cycle time, minimize resource utilization (subject to constraints)<br />
<br />
== Simulating Result ==<br />
* "Changes proposed to business processes can be simulated"<br />
* "'''Sensitivity''' of making the changes on the ultimate objectives can be examined and quantified, reducing the risk of actual implementation"<br />
<br />
; Changes<br />
* adding, deleting, and modifying processes<br />
* process times<br />
* resources required<br />
* schedules<br />
* skill levels<br />
* budgets<br />
; Performance objectives<br />
* throughput<br />
* costs<br />
* inventories<br />
* resources/capital utilization<br />
* cash flow<br />
* waste<br />
<br />
== Uncertaintiy ==<br />
In BPM:<br />
* simulation: a way to understand and communicate the uncertainty related to making the changes,<br />
* optimization: provides the way to manage that uncertainty<br />
<br />
== Academic Approaches ==<br />
# Stochastic approximation (gradient-based approaches)<br />
# (sequential) response surface methodology<br />
# random search<br />
# sample path optimization (also known as stochastic counterpart)<br />
<br />
== Pragmatic Approach ==<br />
* Commercial simulation software employs metaheuristics <br />
https://editor.signavio.com/p/model/a7cda803cb4d48e894671100c040e892/png?inline&authkey=2dc1cf9f7f9d3bdcb96671c66fe0911b6c545adf282da7443bf344a9271822&.png<br />
<br />
<br />
== Evolutionary Approaches == <br />
* commercial simulation uses evolutionary approaches<br />
* Evolutionary approaches: builds and evolves population of solutions<br />
* e.g. '''Genetic Algorithms''' and '''Scatter Search'''.<br />
<br />
* a simulation model can be thought of as a<br />
“mechanism that turns input parameters into output performance measures” (Law and Kelton, 1991)<br />
* simulation model is a function that evaluates the merit of a set of specifications, typically represented as set of values<br />
* Looking at simulation model as a function encouraged family of approaches to optimize simulations based on '''response surfaces''' and '''metamodels'''.<br />
<br />
== Constraints ==<br />
* speciying constraints is important feature of simulation optimization<br />
* constraints define the feasibility of trial solutions<br />
* specified as mathematical expressions or as logic statements<br />
* usually formulated with input factors or/and responses<br />
<br />
== Constraints and Feasability ==<br />
* constraints in a simulation optimization model depend only on input parameters -> new trial solution can be checked for feasibility before running the simulation<br />
* infeasible trial solution may be discarded or may be mapped to a feasible one when its feasibility depends only on constraints formulated with input parameters<br />
* constraints depend on responses -> feasibility of a solution is not known before running the simulation</div>Bernard Szlachta