'Hang On a Minute': Investigations on the Effects of Delayed Objective Functions in Multiobjective Optimization

We consider a multiobjective optimization scenario in which one or more objective functions may be subject to delays (or longer evaluation durations) relative to the other functions. We motivate this scenario from the viewpoint of experimental optimization problems, and derive several simple strategies for dealing with population and/or archive updates under these conditions. These are embedded in a ranking-based EMO algorithm and tested on the WFG test problems augmented with delayed objective(s). Results indicate that good performance can be achieved when the most recently generated solutions are submitted for evaluation on the delayed objective functions, and missing objective values are approximated using a fitness inheritance-based approach. Also, in general one should wait for all evaluations to complete before resuming search if the delay is short, while a non-waiting strategy should be preferred for longer delays.

[1]  Kalyanmoy Deb,et al.  Simulated Binary Crossover for Continuous Search Space , 1995, Complex Syst..

[2]  Stef van Buuren,et al.  Multiple imputation of discrete and continuous data by fully conditional specification , 2007 .

[3]  Joshua D. Knowles,et al.  On-Line Purchasing Strategies for an Evolutionary Algorithm Performing Resource-Constrained Optimization , 2010, PPSN.

[4]  Andrew Lewis,et al.  Asynchronous Multi-Objective Optimisation in Unreliable Distributed Environments , 2009 .

[5]  Ofer M. Shir,et al.  Enhancing Decision Space Diversity in Evolutionary Multiobjective Algorithms , 2009, EMO.

[6]  Gregory Hornby,et al.  ALPS: the age-layered population structure for reducing the problem of premature convergence , 2006, GECCO.

[7]  Kalyanmoy Deb,et al.  Non-linear Dimensionality Reduction Procedures for Certain Large-Dimensional Multi-objective Optimization Problems: Employing Correntropy and a Novel Maximum Variance Unfolding , 2007, EMO.

[8]  Thomas Philip Runarsson,et al.  Constrained Evolutionary Optimization by Approximate Ranking and Surrogate Models , 2004, PPSN.

[9]  Lothar Thiele,et al.  Defining and Optimizing Indicator-Based Diversity Measures in Multiobjective Search , 2010, PPSN.

[10]  Ofer M. Shir,et al.  Quantum control experiments as a testbed for evolutionary multi-objective algorithms , 2012, Genetic Programming and Evolvable Machines.

[11]  Salvatore Greco,et al.  Evolutionary Multi-Criterion Optimization , 2011, Lecture Notes in Computer Science.

[12]  Joshua D. Knowles,et al.  On Handling Ephemeral Resource Constraints in Evolutionary Search , 2013, Evolutionary Computation.

[13]  Joshua D. Knowles,et al.  ParEGO: a hybrid algorithm with on-line landscape approximation for expensive multiobjective optimization problems , 2006, IEEE Transactions on Evolutionary Computation.

[14]  Robert E. Smith,et al.  Fitness inheritance in genetic algorithms , 1995, SAC '95.

[15]  Richard Allmendinger,et al.  Tuning evolutionary search for closed-loop optimization , 2012 .

[16]  Joshua D. Knowles,et al.  Policy learning in resource-constrained optimization , 2011, GECCO '11.

[17]  Kalyanmoy Deb,et al.  A combined genetic adaptive search (GeneAS) for engineering design , 1996 .

[18]  Michael T. M. Emmerich,et al.  Single- and multiobjective evolutionary optimization assisted by Gaussian random field metamodels , 2006, IEEE Transactions on Evolutionary Computation.

[19]  R. Lyndon While,et al.  A review of multiobjective test problems and a scalable test problem toolkit , 2006, IEEE Transactions on Evolutionary Computation.

[20]  Eckart Zitzler,et al.  Objective Reduction in Evolutionary Multiobjective Optimization: Theory and Applications , 2009, Evolutionary Computation.

[21]  Andy J. Keane,et al.  Multi-Objective Optimization Using Surrogates , 2010 .

[22]  Robert Schaefer.  Parallel Problem Solving from Nature - PPSN XI, 11th International Conference, Kraków, Poland, September 11-15, 2010. Proceedings, Part II , 2010, PPSN.

[23]  Joshua D. Knowles,et al.  Closed-loop, multiobjective optimization of analytical instrumentation: gas chromatography/time-of-flight mass spectrometry of the metabolomes of human serum and of yeast fermentations. , 2005, Analytical chemistry.

[24]  Joshua D. Knowles,et al.  Efficient discovery of anti-inflammatory small molecule combinations using evolutionary computing , 2011, Nature chemical biology.

[25]  Andrew Lewis,et al.  Asynchronous multiple objective particle swarm optimisation in unreliable distributed environments , 2008, 2008 IEEE Congress on Evolutionary Computation (IEEE World Congress on Computational Intelligence).