Juan F. R. Herrera

On a Hybrid MPI-Pthread Approach for Simplicial Branch-and-Bound

We investigate models that efficiently map branch-and-bound algorithms on a distributed computer architecture using a case of multidimensional Lipschitz Global Optimization. A combination of MPI and Pthreads is studied: MPI for distributed computation (inter-node) and Pthreads for multicore computation (intra-node). That model adapts the algorithm to the characteristics of the architecture at hand with an increasing number of nodes. Dynamic load balancing is performed in intra-node space through dynamic generation of threads. Results show performance improvements compared to OpenMP and MPI versions used in previous work.

On Simplicial Longest Edge Bisection in Lipschitz Global Optimization

Simplicial subsets are popular in branch-and-bound methods for Global Optimization. Longest Edge Bisection is a convenient way to divide a simplex. When the number of dimensions is greater than two, irregular simplices (not all edges have the same length) may appear with more than one longest edge. In these cases, the first longest edge is usually selected. We study the impact of other selection rule of the longest edge to be bisected next on the development of a branch-and-bound algorithm to solve multidimensional Lipschitz Global Optimization instances. Experiments show a significant reduction in the number of evaluated simplices for most of the test problems.

Dynamic and hierarchical load-balancing techniques applied to parallel branch-and-bound methods

Most Lipschitzian Global Optimization algorithms perform an exhaustive search using a branch-and-bound (B&B) scheme. The question is how to run multi-dimensional Lipschitz Global Optimization in parallel, such that the implementation depending on the used platform is efficient. Previous work shows a parallel version developed for multicore nodes with two levels of parallelism: intra-node and inter-node. On intra-node level, one can perform dynamic load balancing by generating threads dynamically. Threads end when they complete their assigned work. The inter-node level carries out a static load balancing using MPI. In general, algorithm design depends on the characteristics of problems to be solved. There are several ways to improve performance of general parallel B&B algorithms. Specifically, we are interested in how to apply them to parallel Lipschitz Global Optimization algorithms. Operations like selecting the next subproblem to be evaluated become critical in parallel B&B schemes. We study Depth and Hybrid (Best-Depth) options as selection criterion. Previous work, using only MPI or OpenMP, discard not only the broadcasting of the best found upper bound of the solution due to its high average cost/performance but also the use of dynamic load balancing. Here we check how broadcasting the upper bound affects the developed MPI-Pthreads algorithm. Additionally, we study how to perform dynamic load balancing at inter-node level. Experimental results show which designs perform better on which type of instances for the used computational architecture.

Heuristics for Longest Edge Selection in Simplicial Branch and Bound

Simplicial partitions are suitable to divide a bounded area in branch and bound. In the iterative refinement process, a popular strategy is to divide simplices by their longest edge, thus avoiding needle-shaped simplices. A range of possibilities arises in higher dimensions where the number of longest edges in a simplex is greater than one. The behaviour of the search and the resulting binary search tree depend on the selected longest edge. In this work, we investigate different rules to select a longest edge and study the resulting efficiency of the branch and bound algorithm.

Data Parallelism in Traffic Control Tables with Arrival Information

Traffic lights can be controlled dynamically through rules reacting on the number of waiting vehicles at each light. A rule can be captured by a so-called Traffic Control Table TCT. The Value Iteration method from Stochastic Dynamic Programming has been used for simple networks to derive a TCT. This work studies the generation of a TCT-based rule that takes the arrival information of new vehicles into account. The question is how to generate such a table for simple intersections or a network of these. The generation is particularly difficult due to the computational work involved in the Value Iteration process. The problem is formulated as a Markov Decision Process and the parallelization of the Value Iteration method for this problem is discussed. We are specifically interested in exploiting the structure of the problem for simple infrastructures, with only a few traffic lanes, using a parallel algorithm.

A threaded approach of the quadratic bi-blending algorithm

Blending algorithms aim for solving the problem of determining the mixture of raw materials in order to obtain a cheap and feasible recipe with the smallest number of raw materials. An algorithm that solves this problem for two products, where available raw material is limited, has two phases. The first phase is a simplicial branch-and-bound algorithm which determines, for a given precision, a Pareto set of solutions of the bi-blending problem as well as a subspace of the initial space where better feasible solutions (with more precision) can be found. The second phase basically consists in an exhaustive reduction of the mentioned subspace by deleting simplicial subsets that do not contain solutions. This second phase is useful for future refinement of the solutions. Previous work only focused on the first phase neglecting the second phase due to computational burden. With this in mind, we study the parallelization of the different phases of the sequential bi-blending algorithm and focus on the most time consuming phase, analyzing the performance of several strategies.

1076TiPRANDOMIZED PHASE II TRIAL WITH DENDRITIC CELL (DC) IMMUNOTHERAPY IN PATIENTS WITH COLORECTAL CARCINOMA AND LIVER METASTASIS FOLLOWING COMPLETE RESECTION AND ADJUVANT CHEMOTHERAPY

ABSTRACT Background: Cellular immunotherapy with DC has well-established anti-tumor activity, as confirmed by Sipuleucel-T, the first approved treatment based in this strategy. Preclinical and clinical data indicate that efficacy of DC immunotherapy is maximized when used in the setting of minimal residual disease. Our study explores the efficacy of an autologous DC vaccine loaded with self-tumor antigens that we developed in a previous study (Alfaro, J Immunology 2011) in patients with colorectal cancer that have undergone complete resection of hepatic metastasis and standard adjuvant therapy. Trial design: In this randomized phase II study, patients with colorectal carcinoma with hepatic metastasis that have undergone standard treatment, including complete surgical resection and standard adjuvant chemotherapy, are randomized to receive DC vaccine or to observation. The two-cycle vaccination protocol includes the following strategies: a) pretreatment with cyclophosphamide to decrease regulatory T cells; b) maturation and activation of DC with TNF-alpha, interferon-alpha and poly I:C, a potent inducer of type I interferon; c) use of autologous tumor from resected liver metastasis as antigenic source, to include antigens that are exclusive of tumor cells; and d) administration of daily intradermal vaccines during four consecutive days in 2 cycles every 4 weeks. Our aim is to replicate the immune response observed during an acute viral infection in terms of activation signals and persistence of antigens in lymph nodes. The main objective is progression-free survival. Thirty-six patients will be included, allowing to detect a HR = 1.75 (alpha error = 0.2, beta error = 0.35). Secondary objectives are: safety, overall survival and immunologic response (in vitro lymphocyte responses against tumor antigens; delayed hypersensitivity reactions; induction of tumor antibody responses; DC activation parameters including IL-12 and IL-6 production and expression of CD80, CD83, CD86, B7-H1, B7-H4 and B7-DC; assessment of DC maturation by expression of pro-inflammatory cytokines; and DC migration) (ClinicalTrials.gov Identifier: NCT01348256). This abstract was accepted and previously presented at the 2014 ASCO Annual Meeting Chicago, June 2014 (TPS3129). Disclosure: All authors have declared no conflicts of interest.