Automatic generation and exploitation of related problems in genetic programming

Krawiec, Krzysztof; Wieloch, Bartosz

doi:10.1109/cec.2010.5586120

Cited by 13 publications

(6 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…With the above in mind, in this paper, we demonstrate that the practicality of population-based bi-level optimization can be significantly enhanced simply by incorporating the novel concept of evolutionary multitasking into the search process. As has been demonstrated in [19,22], evolutionary multitasking provides the scope for accelerating convergence to near optimal solutions of multiple optimization tasks at once, simply by harnessing the latent complementarities among them (when available). In particular, the true power of implicit parallelism of population-based search is unleashed by combining the design spaces corresponding to different tasks into a unified pool of genetic material, thereby facilitating implicit information exchange among them in the form of encoded genetic material.…”

Section: Introductionmentioning

confidence: 99%

Evolutionary multitasking in bi-level optimization

Gupta

Mańdziuk

Ong

2015

Complex Intell. Syst.

View full text Add to dashboard Cite

Evolutionary multitasking has recently emerged as an effective means of facilitating implicit genetic transfer across different optimization tasks, thereby potentially accelerating convergence characteristics for multiple tasks at once. A natural application of the paradigm is found to arise in the area of bi-level programming wherein an upper level optimization problem must take into consideration a nested lower level problem. Thus, while tackling instances of bi-level optimization, a significant challenge surfaces from the fact that multiple upper level candidate solutions are to be analyzed at the same time by inferring the corresponding optimum response from the lower level. Thus, the process of bi-level optimization often becomes exorbitantly time consuming, especially in the case of real-world instances involving expensive objective function evaluations. Accordingly, the significance of this paper lies in showcasing that the practicality of population-based bi-level optimization can be considerably enhanced by simply incorporating the novel concept of evolutionary multitasking into the search process. As a result, it becomes possible to process multiple lower level optimization tasks concurrently, thereby facilitating the exploitation of underlying commonalities among them. To demonstrate the implications of our proposal, we present computational experiments on some synthetic benchmark B Yew-Soon Ong

show abstract

Section: Introductionmentioning

confidence: 99%

Evolutionary multitasking in bi-level optimization

Gupta

Mańdziuk

Ong

2015

Complex Intell. Syst.

View full text Add to dashboard Cite

show abstract

“…Returning to Algorithm 1, note that a rank-based fitness calculation scheme, as has been described in [17,25] for inferring the overall fitness of an individual in a multitasking environment, is employed in the generational selection step of the MFEA. Further, bear in mind that the skill factors of all individuals in the initial population, as they clearly cannot be assigned via the inductive process of imitation, can either be assigned randomly (while ensuring uniform representation for all constitutive optimization tasks) or by the exhaustive evaluation procedure considered in [17].…”

Section: An Evolutionary Methodology For Multitask Optimizationmentioning

confidence: 99%

Evolutionary Multitasking: A Computer Science View of Cognitive Multitasking

2016

View full text Add to dashboard Cite

The human mind possesses the most remarkable ability to perform multiple tasks with apparent simultaneity. In fact, with the present-day explosion in the variety and volume of incoming information streams that must be absorbed and appropriately processed, the opportunity, tendency, and (even) the need to multitask are unprecedented. Thus, it comes as little surprise that the pursuit of intelligent systems and algorithms that are capable of efficient multitasking is rapidly gaining importance among contemporary scientists who are faced with the increasing complexity of real-world problems. To this end, the present paper is dedicated to a detailed exposition on a so-far underexplored characteristic of population-based search algorithms, i.e., their inherent ability (much like the human mind) to handle multiple optimization tasks at once. We present a simple evolutionary methodology capable of cross-domain multitask optimization in a unified genotype space and show that there exist many potential benefits of its application in practical domains. Most notably, it is revealed that multitasking enables one to automatically leverage upon the underlying commonalities between distinct optimization tasks, thereby providing the scope for considerably improved performance in real-world problem solving.

show abstract

“…We verify whether the semantic properties of LGX influence search efficiency by solving univariate symbolic regression problems shown in Table 2, taken from [4], using instruction set {+, −, ×, /, x}. Semantics is defined as a vector of values returned by a program for 20 fitness cases distributed equidistantly in the interval [−1, 1].…”

Section: The Experimentsmentioning

confidence: 99%

Locally geometric semantic crossover

Krawiec

Pawlak

2012

Proceedings of the 14th Annual Conference Companion on Genetic and Evolutionary Computation

Self Cite

View full text Add to dashboard Cite

We propose Locally Geometric Crossover (LGX) for genetic programming. For a pair of homologous loci in the parent solutions, LGX finds a semantically intermediate procedure from a previously prepared library, and uses it as replacement code. The experiments involving six symbolic regression problems show significant increase in search performance when compared to standard subtree-swapping crossover and other control methods. This suggests that semantically geometric manipulations on subprograms propagate to entire programs and improve their fitness. Keywordsgenetic programming, program semantics, semantic crossover MOTIVATION AND THE METHODStandard search operators used in genetic programming (GP) are blind to program semantic. For them, programs are purely symbolic structures, where opcodes associated with particular instructions have no particular meaning. Given that knowledge of semantics is one of factors that makes human programming so effective, equipping GP algorithms in some semantic-aware extensions could have positive impact on their performance.We devise a method that makes GP alert to certain semantic aspects of programs. For this, we assume that fitness function is a metric || || that captures the divergence (error) between program output and some known desired output, which is typically the case in GP. Consequently, the fitness landscape is a convex surface spanned over the space of vectors that hold program outputs.Convexity allows designing recombination operators that are likely to yield offspring of good quality [6]. We exploit this property locally, on the level of program parts, rather than entire programs. The method operates in two phases. First, a library L of short procedures of height at most h is generated. Next, we calculate the semantics s(p) of every procedure p ∈ L, meant here as the vector of outcomes Copyright is held by the author/owner(s). GECCO'12 Companion, July 7-11, 2012, Philadelphia, PA, USA. ACM 978-1-4503-1178-6/12/07. it produces for the fitness cases. For any two procedures p 1, p2 that have the same semantics (||s(p1) − s(p2)|| = 0) we discard the longer one to avoid redundancy.In the second phase, the library is used by the proposed locally geometric crossover (LGX) during a GP run. Given two parent programs p 1 and p2, LGX first identifies the structurally common region [7], i.e., the set of tree node locations that occur in both parents. Then LGX appoints a single random locus in the common region as a crossover point. To reduce bloat, this follows the same rules as in standard GP [2]: with probability 0.1, the locus is a leaf in the common region, otherwise an internal node.Let p 1 and p 2 denote the subtrees rooted in the drawn locus in p 1 and p2, respectively. We calculate their semantics, s(p 1 ) and s(p 2 ), and determine the midpoint between them in the semantic space: s m = (s(p 1 ) + s(p 2 ))/2. This point represents the semantics of a hypothetical procedure p : s m = s(p) which, if inserted into parents at the appointed locus, would make the resulting offs...

show abstract

Automatic generation and exploitation of related problems in genetic programming

Cited by 13 publications

References 11 publications

Evolutionary multitasking in bi-level optimization

Evolutionary multitasking in bi-level optimization

Evolutionary Multitasking: A Computer Science View of Cognitive Multitasking

Locally geometric semantic crossover

Contact Info

Product

Resources

About