GENESYS: A novel evolutionary program synthesis tool with continuous optimization

GENET and its extended model EGENET are artificial neural networks to efficiently solve finite constraint satisfaction problems such as the car-sequencing problems. Both models use the min-conflict heuristic to update every finite-domain variable for finding local minima, and then apply heuristic learning rule(s) to escape the local minima not representing solution(s). Since continuous and finite domains are completely different, researchers seldom considered to apply the EGENET approach to solve continuous constrained optimization problems. We consider an interesting proposal to modify the original EGENET model with the minimal effort for solving continuous constrained optimization problems. Our proposal immediately opens up new directions for studying many possible ways to approximate continuous domains using modified finite-domain solvers. Moreover, the preliminary benchmarks of our prototypes on some graph layout problems as practical examples demonstrated some advantages of our proposal which prompts for further investigation.

In optimization and machine learning, the divide between discrete and continuous problems and methods is deep and persistent. We attempt to remove this distinction by training neural network autoencoders that embed discrete candidate solutions in continuous latent spaces. This allows us to take advantage of state-of-the-art continuous optimization methods for solving discrete optimization problems, and mitigates certain challenges in discrete optimization, such as design of bias-free search operators. In the experimental part, we consider program synthesis as the special case of combinatorial optimization. We train an autoencoder network on a large sample of programs in a problem-agnostic, unsupervised manner, and then use it with an evolutionary continuous optimization algorithm (CMA-ES) to map the points from the latent space to programs. We propose also a variant in which semantically similar programs are more likely to have similar embeddings. Assessment on a range of benchmarks in two domains indicates the viability of this approach and the usefulness of involving program semantics.

A competitive variable-fidelity surrogate-assisted CMA-ES algorithm using data mining techniques

We present a method for example-guided synthesis of functional programs over recursive data structures. Given a set of input-output examples, our method synthesizes a program in a functional language with higher-order combinators like map and fold. The synthesized program is guaranteed to be the simplest program in the language to fit the examples. Our approach combines three technical ideas: inductive generalization, deduction, and enumerative search. First, we generalize the input-output examples into hypotheses about the structure of the target program. For each hypothesis, we use deduction to infer new input/output examples for the missing subexpressions. This leads to a new subproblem where the goal is to synthesize expressions within each hypothesis. Since not every hypothesis can be realized into a program that fits the examples, we use a combination of best-first enumeration and deduction to search for a hypothesis that meets our needs. We have implemented our method in a tool called λ2, and we evaluate this tool on a large set of synthesis problems involving lists, trees, and nested data structures. The experiments demonstrate the scalability and broad scope of λ2. A highlight is the synthesis of a program believed to be the world's earliest functional pearl.

In combinatorial solution spaces Iterated Local Search (ILS) turns out to be exceptionally successful. The question arises: is ILS also capable of improving the optimization process in continuous solution spaces? To demonstrate that hybridization leads to powerful techniques in continuous domains, we introduce a hybrid meta-heuristic that integrates Powell’s direct search method. It combines direct search with elements from population based evolutionary optimization. The approach is analyzed experimentally on a set of well known test problems and compared to a state-of-the-art technique, i.e., a restart variant of the Covariance Matrix Adaptation Evolution Strategy with increasing population sizes (G-CMA-ES). It turns out that the population-based Powell-ILS is competitive to the CMA-ES, in some cases even significantly faster and behaves more robust than the pure strategy of Powell in multimodal fitness landscapes. Further experiments on the perturbation mechanism, population sizes, and problems with noise complete the analysis of the hybrid methodology and lead to parameter recommendations.

Population-based approaches have given us new search strategies and ideas in order to solve optimization problems. Usually, these methods are based on the performance carried out by a finite number of agents, which by the interaction between them they evolve and work all over the search space. Also, it is well-known that the correct employment of parameter values in this kind of method can positively impact their performance and behavior. In this context, the present work focuses on the design of a hybrid architecture which smartly balances the population size on run-time. In order to smartly balance and control the population size, a modular approach, named Linear Modular Population Balancer (LMPB), is proposed. The main ideas behind the designed architecture include the solving strategy behind a population-based metaheuristic, the influence of learning components based on multiple statistical modeling methods which transform the dynamic data generated into knowledge, and the possibilities to tackle both discrete and continuous optimization problems. In this regard, three modules are proposed for LMPB, which concern tasks such as the management of the population-based algorithm, parameter setting, probabilities, learning methods, and selection mechanism for the population size to employ. In order to test the viability and effectiveness of our proposed approach, we solve a set of well-known benchmark functions and the multidimensional knapsack problem (MKP). Additionally, we illustrate promising solving results, compare them against state-of-the-art methods which have proved to be good options for solving optimization problems, and give solid arguments for future work in the necessity to keep evolving this type of proposed architecture.

Program verification is the task of automatically generating proofs for a program’s compliance with a given specification. Program synthesis is the task of automatically generating a program that meets a given specification. Both program verification and program synthesis can be viewed as search problems, for proofs and programs, respectively. For these search problems, we present approaches based on user-provided insights in the form of templates. Templates are hints about the syntactic forms of the invariants and programs, and help guide the search for solutions. We show how to reduce the template-based search problem to satisfiability solving, which permits the use of off-the-shelf solvers to efficiently explore the search space. Template-based approaches have allowed us to verify and synthesize programs outside the abilities of previous verifiers and synthesizers. Our approach can verify and synthesize difficult algorithmic textbook programs (e.g., sorting and dynamic programming-based algorithms) and difficult arithmetic programs.

Randomized direct search algorithms for continuous domains, such as evolution strategies, are basic tools in machine learning. They are especially needed when the gradient of an objective function (e.g., loss, energy, or reward function) cannot be computed or estimated efficiently. Application areas include supervised and reinforcement learning as well as model selection. These randomized search strategies often rely on normally distributed additive variations of candidate solutions. In order to efficiently search in non-separable and ill-conditioned landscapes the covariance matrix of the normal distribution must be adapted, amounting to a variable metric method. Consequently, covariance matrix adaptation (CMA) is considered state-of-the-art in evolution strategies. In order to sample the normal distribution, the adapted covariance matrix needs to be decomposed, requiring in general Θ(n 3) operations, where n is the search space dimension. We propose a new update mechanism which can replace a rank-one covariance matrix update and the computationally expensive decomposition of the covariance matrix. The newly developed update rule reduces the computational complexity of the rank-one covariance matrix adaptation to Θ(n 2) without resorting to outdated distributions. We derive new versions of the elitist covariance matrix adaptation evolution strategy (CMA-ES) and the multi-objective CMA-ES. These algorithms are equivalent to the original procedures except that the update step for the variable metric distribution scales better in the problem dimension. We also introduce a simplified variant of the non-elitist CMA-ES with the incremental covariance matrix update and investigate its performance. Apart from the reduced time-complexity of the distribution update, the algebraic computations involved in all new algorithms are simpler compared to the original versions. The new update rule improves the performance of the CMA-ES for large scale machine learning problems in which the objective function can be evaluated fast.

In this study, we investigate the problem of min-max continuous optimization in a black-box setting $\min_{x} \max_{y}f(x,y)$. A popular approach updates $x$ and $y$ simultaneously or alternatingly. However, two major limitations have been reported in existing approaches. (I) As the influence of the interaction term between $x$ and $y$ (e.g., $x^\mathrm{T} B y$) on the Lipschitz smooth and strongly convex-concave function $f$ increases, the approaches converge to an optimal solution at a slower rate. (II) The approaches fail to converge if $f$ is not Lipschitz smooth and strongly convex-concave around the optimal solution. To address these difficulties, we propose minimizing the worst-case objective function $F(x)=\max_{y}f(x,y)$ directly using the covariance matrix adaptation evolution strategy, in which the rankings of solution candidates are approximated by our proposed worst-case ranking approximation (WRA) mechanism. Compared with existing approaches, numerical experiments show two important findings regarding our proposed method. (1) The proposed approach is efficient in terms of $f$-calls on a Lipschitz smooth and strongly convex-concave function with a large interaction term. (2) The proposed approach can converge on functions that are not Lipschitz smooth and strongly convex-concave around the optimal solution, whereas existing approaches fail.

Optimization problems with continuous variables are very often in real-life scenarios. This is why several studies have focused on the use of metaheuristics to solve these problems, known as continuous optimization problems. This chapter focuses on the robust population-based metaheuristic Variable Mesh Optimization (VMO). We review the fundamental aspects of VMO for continuous optimization. Moreover, we analyze the variants of VMO that have been proposed to solve multimodal problems, i.e. those with multiple optima at several search subspaces. The chapter uses a homogeneous algorithmic framework to describe VMO and its variants with niching strategies for multimodal problems.KeywordsMetaheuristicsVariable mesh optimizationContinuous optimization problemsMultimodal optimization problemsNiching

Research into ant colony algorithms for solving continuous optimization problems forms one of the most significant and promising areas in swarm computation. Although traditional ant algorithms are designed for combinatorial optimization, they have shown great potential in solving a wide range of optimization problems, including continuous optimization. Aimed at solving continuous problems effectively, this paper develops a novel ant algorithm termed “continuous orthogonal ant colony” (COAC), whose pheromone deposit mechanisms would enable ants to search for solutions collaboratively and effectively. By using the orthogonal design method, ants in the feasible domain can explore their chosen regions rapidly and efficiently. By implementing an “adaptive regional radius” method, the proposed algorithm can reduce the probability of being trapped in local optima and therefore enhance the global search capability and accuracy. An elitist strategy is also employed to reserve the most valuable points. The performance of the COAC is compared with two other ant algorithms for continuous optimization — API and CACO by testing seventeen functions in the continuous domain. The results demonstrate that the proposed COAC algorithm outperforms the others.

Evolutionary algorithms and elliptical copulas applied to continuous optimization problems

Surrogate functions are often employed to reduce the number of objective function evaluations in a continuous optimization. However, their effects have seldom been investigated theoretically. This paper analyzes the effect of a surrogate function in the information-geometric optimization (IGO) framework, which includes as an algorithm instance a variant of the covariance matrix adaptation evolution strategy—a widely used solver for black-box continuous optimization. We derive a sufficient condition on the surrogate function for the parameter update in the IGO algorithms to point to a descent direction of the objective function expected over the search distribution. The condition is expressed in terms of three measures of correlation between the objective function and the surrogate function. Our result constitutes a partial justification for the use of a surrogate function in IGO algorithms.

Causal discovery by continuous optimization with weighted superstructure.

Optimum implementation of non-conventional wells allows us to increase considerably hydrocarbon recovery. By considering the high drilling cost and the potential improvement in well productivity, well placement decision is an important issue in field development. Considering complex reservoir geology and high reservoir heterogeneities, stochastic optimization methods are the most suitable approaches for optimum well placement. This paper proposes an optimization methodology to determine optimal well location and trajectory based upon the Covariance Matrix Adaptation-Evolution Strategy (CMA-ES) which is a variant of Evolution Strategies recognized as one of the most powerful derivative-free optimizers for continuous optimization. To improve the optimization procedure, two new techniques are investigated: (1). Adaptive penalization with rejection is developed to handle well placement constraints. (2). A meta-model, based on locally weighted regression, is incorporated into CMA-ES using an approximate ranking procedure. Therefore, we can reduce the number of reservoir simulations, which are computationally expensive. Several examples are presented. Our new approach is compared with a Genetic Algorithm incorporating the Genocop III technique. It is shown that our approach outperforms the genetic algorithm: it leads in general to both a higher NPV and a significant reduction of the number of reservoir simulations. ECMOR XII – 12 th European Conference on the Mathematics of Oil Recovery