{"id":9735,"date":"2025-03-21T16:09:10","date_gmt":"2025-03-21T16:09:10","guid":{"rendered":"https:\/\/meetings.informs.org\/wordpress\/analytics2025\/?page_id=9735"},"modified":"2026-03-05T14:27:51","modified_gmt":"2026-03-05T14:27:51","slug":"iaaa","status":"publish","type":"page","link":"https:\/\/meetings.informs.org\/wordpress\/analytics2025\/iaaa\/","title":{"rendered":"IAAA Finalists"},"content":{"rendered":"<!--themify_builder_content-->\n<div id=\"themify_builder_content-9735\" data-postid=\"9735\" class=\"themify_builder_content themify_builder_content-9735 themify_builder tf_clear\">\n                    <div  data-css_id=\"y990649\" data-lazy=\"1\" class=\"module_row themify_builder_row fullwidth_row_container tb_y990649 tb_first tf_w\">\n                        <div class=\"row_inner col_align_top tb_col_count_1 tf_box tf_rel\">\n                        <div  data-lazy=\"1\" class=\"module_column tb-column col-full tb_p5xb445 first\">\n                    <!-- module text -->\n<div  class=\"module module-text tb_dw32055   \" data-lazy=\"1\">\n        <div  class=\"tb_text_wrap\">\n        <h2>Track: INFORMS Prizes &#8211; IAAA Finalists<\/h2>    <\/div>\n<\/div>\n<!-- \/module text --><!-- module text -->\n<div  class=\"module module-text tb_8yiu590   \" data-lazy=\"1\">\n        <div  class=\"tb_text_wrap\">\n        <p>The IAAA Finalists will present at the conference on Tuesday, April 8 in the INFORMS Prizes Track from 9:10am \u2013 10:00am EST. Judges will then review the rank and announce the winner at the INFORMS Analytics Society luncheon.<\/p>    <\/div>\n<\/div>\n<!-- \/module text -->        <\/div>\n                        <\/div>\n        <\/div>\n                        <div  data-lazy=\"1\" class=\"module_row themify_builder_row tb_qsmi272 tf_w\">\n                        <div class=\"row_inner col_align_top tb_col_count_1 tf_box tf_rel\">\n                        <div  data-lazy=\"1\" class=\"module_column tb-column col-full tb_l69b273 first\">\n                    <!-- module accordion -->\n<div  class=\"module module-accordion tb_94pc306 \" data-behavior=\"toggle\" data-lazy=\"1\">\n    \n    <ul class=\"ui module-accordion   tb_default_color\">\n            <li>\n            <div class=\"accordion-title tf_rel\">\n                <a href=\"#acc-94pc306-0\" class=\"tb_title_accordion\" aria-controls=\"acc-94pc306-0-content\" aria-expanded=\"false\">\n                    <i class=\"accordion-icon\"><svg  class=\"tf_fa tf-ti-plus\" aria-hidden=\"true\"><use href=\"#tf-ti-plus\"><\/use><\/svg><\/i>                    <i class=\"accordion-active-icon tf_hide\"><svg  class=\"tf_fa tf-ti-minus\" aria-hidden=\"true\"><use href=\"#tf-ti-minus\"><\/use><\/svg><\/i>                    <span class=\"accordion-title-wrap\">Spot Market Pricing on Amazon Freight<\/span>                <\/a>\n            <\/div><!-- .accordion-title -->\n            <div id=\"acc-94pc306-0-content\" data-id=\"acc-94pc306-0\" aria-hidden=\"true\" class=\"accordion-content tf_hide tf_clearfix\">\n                        <div  data-lazy=\"1\" class=\"module_subrow themify_builder_sub_row tf_w col_align_top tb_col_count_1 tb_53dd971\">\n                <div  data-lazy=\"1\" class=\"module_column sub_column tb_z4eo971 first\">\n                    <!-- module text -->\n<div  class=\"module module-text tb_nnvc971   \" data-lazy=\"1\">\n        <div  class=\"tb_text_wrap\">\n        <p><b><span data-contrast=\"auto\">Authors:<\/span><\/b><span data-contrast=\"auto\"> Adam N. Elmachtoub , Goutam Kumar, Roger Lederman<\/span><span data-ccp-props=\"{&quot;335559685&quot;:360}\">\u00a0<\/span><\/p>\n<p><b><span data-contrast=\"auto\">Affiliation:<\/span><\/b><span data-contrast=\"auto\"> Amazon.<\/span><span data-ccp-props=\"{&quot;335559685&quot;:360}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">Amazon Freight (AF) is a business unit at Amazon that moves freight for other<\/span><span data-ccp-props=\"{&quot;335559685&quot;:360}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">shippers using Amazon\u2019s internal middle mile network. For this business, external shippers<\/span><span data-ccp-props=\"{&quot;335559685&quot;:360}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">use a website to request a price quote for an origin-destination pair, get prices instantly,<\/span><span data-ccp-props=\"{&quot;335559685&quot;:360}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">and decide which option to book. AF provides prices that varies across lead time options.<\/span><span data-ccp-props=\"{&quot;335559685&quot;:360}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">The execution costs and demand can vary widely across different quotes due to internal cost<\/span><span data-ccp-props=\"{&quot;335559685&quot;:360}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">structure and overall freight market. The goal for AF is to set prices for each quote in order<\/span><span data-ccp-props=\"{&quot;335559685&quot;:360}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">to maximize a business objective. Our solution utilizes a combination of parameteric and<\/span><span data-ccp-props=\"{&quot;335559685&quot;:360}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">non-parametric modeling with price optimization. We first use quote features to segment<\/span><span data-ccp-props=\"{&quot;335559685&quot;:360}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">the market into disjoint groups, employing the Market Segmentation Tree (MST) algorithm<\/span><span data-ccp-props=\"{&quot;335559685&quot;:360}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">which creates a binary tree based on differences in choice behavior. Within each leaf, we fit<\/span><span data-ccp-props=\"{&quot;335559685&quot;:360}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">a reference-price-effects Multinomial (MNL) choice model that is amenable to fast pricing<\/span><span data-ccp-props=\"{&quot;335559685&quot;:360}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">heuristics. We conducted live A\/B experiments that shows our new framework significantly<\/span><span data-ccp-props=\"{&quot;335559685&quot;:360}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">improves profit, revenue, and accuracy. The model has been in production for about a year.<\/span><span data-ccp-props=\"{&quot;335559685&quot;:360}\">\u00a0<\/span><\/p>    <\/div>\n<\/div>\n<!-- \/module text -->        <\/div>\n                    <\/div>\n                    <\/div><!-- .accordion-content -->\n        <\/li>\n        <\/ul>\n\n<\/div><!-- \/module accordion --><!-- module accordion -->\n<div  class=\"module module-accordion tb_q7hq474 \" data-behavior=\"toggle\" data-lazy=\"1\">\n    \n    <ul class=\"ui module-accordion   tb_default_color\">\n            <li>\n            <div class=\"accordion-title tf_rel\">\n                <a href=\"#acc-q7hq474-0\" class=\"tb_title_accordion\" aria-controls=\"acc-q7hq474-0-content\" aria-expanded=\"false\">\n                    <i class=\"accordion-icon\"><svg  class=\"tf_fa tf-ti-plus\" aria-hidden=\"true\"><use href=\"#tf-ti-plus\"><\/use><\/svg><\/i>                    <i class=\"accordion-active-icon tf_hide\"><svg  class=\"tf_fa tf-ti-minus\" aria-hidden=\"true\"><use href=\"#tf-ti-minus\"><\/use><\/svg><\/i>                    <span class=\"accordion-title-wrap\">Semantically Grounded Analytics Engine for Helping Patients Find Doctors and Locations in a Large Healthcare Organization <\/span>                <\/a>\n            <\/div><!-- .accordion-title -->\n            <div id=\"acc-q7hq474-0-content\" data-id=\"acc-q7hq474-0\" aria-hidden=\"true\" class=\"accordion-content tf_hide tf_clearfix\">\n                        <div  data-lazy=\"1\" class=\"module_subrow themify_builder_sub_row tf_w col_align_top tb_col_count_1 tb_kpqd971\">\n                <div  data-lazy=\"1\" class=\"module_column sub_column tb_swpd971 first\">\n                    <!-- module text -->\n<div  class=\"module module-text tb_1ca7971   \" data-lazy=\"1\">\n        <div  class=\"tb_text_wrap\">\n        <p><b><span data-contrast=\"auto\">Authors:<\/span><\/b><span data-contrast=\"auto\"> Mayank Kejriwal*, Hamid Haidarian**, Min-Hsueh Chiu*, Andy Xiang**, Deep Shrestha**, Faizan Javed**<\/span><span data-ccp-props=\"{&quot;335559685&quot;:360}\">\u00a0<\/span><\/p>\n<p><b><span data-contrast=\"auto\">Affiliations:<\/span><\/b><span data-contrast=\"auto\">\u202f*University of Southern California; **Kaiser Permanente Digital<\/span><span data-ccp-props=\"{&quot;335559685&quot;:360}\">\u00a0<\/span><\/p>\n<p><span class=\"TextRun SCXW137975790 BCX8\" lang=\"EN-US\" xml:lang=\"EN-US\" data-contrast=\"auto\"><span class=\"NormalTextRun SCXW137975790 BCX8\">Efficiently finding doctors and locations (FDL) is an important search problem for patients in the healthcare domain, for which traditional analytical and information retrieval approaches tend to be sub-optimal. Legacy FDL systems, based usually on keyword and simple pattern matching, lack\u202f<\/span><span class=\"NormalTextRun SCXW137975790 BCX8\">semantic<\/span><span class=\"NormalTextRun SCXW137975790 BCX8\">\u202ffoundations, preventing them from understanding users\u2019 underlying intent. Through innovations in all three prongs of analytics (descriptive, predictive, and prescriptive), we present a solution to this problem that uses a rich combination of a healthcare knowledge graph and concept-rich ontology. The solution has been implemented as a real-time system for FDL in Kaiser Permanente, a large healthcare organization with <\/span><span class=\"NormalTextRun SCXW137975790 BCX8\">12 million members<\/span><span class=\"NormalTextRun SCXW137975790 BCX8\">, and offers significant performance benefits while <\/span><span class=\"NormalTextRun SCXW137975790 BCX8\">maintaining<\/span><span class=\"NormalTextRun SCXW137975790 BCX8\"> practical needs of data security and privacy, scalability, cost-effectiveness, and backward compatibility with existing search infrastructure<\/span><\/span><\/p>    <\/div>\n<\/div>\n<!-- \/module text -->        <\/div>\n                    <\/div>\n                    <\/div><!-- .accordion-content -->\n        <\/li>\n        <\/ul>\n\n<\/div><!-- \/module accordion --><!-- module accordion -->\n<div  class=\"module module-accordion tb_7e1t878 \" data-behavior=\"toggle\" data-lazy=\"1\">\n    \n    <ul class=\"ui module-accordion   tb_default_color\">\n            <li>\n            <div class=\"accordion-title tf_rel\">\n                <a href=\"#acc-7e1t878-0\" class=\"tb_title_accordion\" aria-controls=\"acc-7e1t878-0-content\" aria-expanded=\"false\">\n                    <i class=\"accordion-icon\"><svg  class=\"tf_fa tf-ti-plus\" aria-hidden=\"true\"><use href=\"#tf-ti-plus\"><\/use><\/svg><\/i>                    <i class=\"accordion-active-icon tf_hide\"><svg  class=\"tf_fa tf-ti-minus\" aria-hidden=\"true\"><use href=\"#tf-ti-minus\"><\/use><\/svg><\/i>                    <span class=\"accordion-title-wrap\">Leveraging AI for AI , Auxiliary Models to sustain a customer facing chatbot<\/span>                <\/a>\n            <\/div><!-- .accordion-title -->\n            <div id=\"acc-7e1t878-0-content\" data-id=\"acc-7e1t878-0\" aria-hidden=\"true\" class=\"accordion-content tf_hide tf_clearfix\">\n                        <div  data-lazy=\"1\" class=\"module_subrow themify_builder_sub_row tf_w col_align_top tb_col_count_1 tb_6lfm971\">\n                <div  data-lazy=\"1\" class=\"module_column sub_column tb_yguz971 first\">\n                    <!-- module text -->\n<div  class=\"module module-text tb_khfk971   \" data-lazy=\"1\">\n        <div  class=\"tb_text_wrap\">\n        <p><strong><span class=\"TextRun SCXW225090491 BCX8\" lang=\"EN-US\" xml:lang=\"EN-US\" data-contrast=\"auto\"><span class=\"NormalTextRun SCXW225090491 BCX8\">Authors:<\/span><\/span><\/strong><span class=\"TextRun SCXW225090491 BCX8\" lang=\"EN-US\" xml:lang=\"EN-US\" data-contrast=\"auto\"><span class=\"NormalTextRun SCXW225090491 BCX8\"> Fabio Dutra Sarti, Safa <\/span><span class=\"NormalTextRun SpellingErrorV2Themed SCXW225090491 BCX8\">Shamout<\/span><span class=\"NormalTextRun SCXW225090491 BCX8\">, Mohan <\/span><span class=\"NormalTextRun SpellingErrorV2Themed SCXW225090491 BCX8\">Mulakkamparambil<\/span><span class=\"NormalTextRun SCXW225090491 BCX8\">, Andres Villegas <\/span><span class=\"NormalTextRun ContextualSpellingAndGrammarErrorV2Themed SCXW225090491 BCX8\">Ceballos<\/span><span class=\"NormalTextRun ContextualSpellingAndGrammarErrorV2Themed SCXW225090491 BCX8\">,\u00a0 <\/span><span class=\"NormalTextRun ContextualSpellingAndGrammarErrorV2Themed SCXW225090491 BCX8\">Antonio<\/span><span class=\"NormalTextRun SCXW225090491 BCX8\"> Henrique, Rafal Orlowski<\/span><\/span><span class=\"EOP SCXW225090491 BCX8\" data-ccp-props=\"{&quot;335559685&quot;:360}\">\u00a0<\/span><\/p>\n<p><span class=\"TextRun SCXW24115224 BCX8\" lang=\"EN-US\" xml:lang=\"EN-US\" data-contrast=\"auto\"><span class=\"NormalTextRun SCXW24115224 BCX8\">This presentation explores the evolution of Scotiabank&#8217;s customer-facing chatbot, from its <\/span><span class=\"NormalTextRun SCXW24115224 BCX8\">initial<\/span><span class=\"NormalTextRun SCXW24115224 BCX8\"> launch and manual maintenance to the automation of key sustainment tasks using machine learning and AI. Scotiabank&#8217;s &#8220;AI for AI&#8221; system, <\/span><span class=\"NormalTextRun SCXW24115224 BCX8\">leveraging<\/span><span class=\"NormalTextRun SCXW24115224 BCX8\"> five proprietary AI models, significantly enhances its award-winning customer support chatbot. This innovative approach automates manual processes, streamlines development, and enables continuous improvement, resulting in reduced workload, enhanced chatbot performance, and increased customer satisfaction. Specifically, it details how auxiliary AI efficiently manages and scales a large-scale chatbot solution, saving thousands of manual work hours. Furthermore, the presentation will highlight the role of data analytics \u2013 through insightful reporting and A\/B testing \u2013 in <\/span><span class=\"NormalTextRun SCXW24115224 BCX8\">optimizing<\/span><span class=\"NormalTextRun SCXW24115224 BCX8\"> customer experience and outcomes. Finally, it will address how Scotiabank has successfully implemented generative AI within a regulated environment.<\/span><\/span><span class=\"EOP SCXW24115224 BCX8\" data-ccp-props=\"{&quot;335559685&quot;:360}\">\u00a0<\/span><\/p>    <\/div>\n<\/div>\n<!-- \/module text -->        <\/div>\n                    <\/div>\n                    <\/div><!-- .accordion-content -->\n        <\/li>\n        <\/ul>\n\n<\/div><!-- \/module accordion --><!-- module accordion -->\n<div  class=\"module module-accordion tb_fojv99 \" data-behavior=\"toggle\" data-lazy=\"1\">\n    \n    <ul class=\"ui module-accordion   tb_default_color\">\n            <li>\n            <div class=\"accordion-title tf_rel\">\n                <a href=\"#acc-fojv99-0\" class=\"tb_title_accordion\" aria-controls=\"acc-fojv99-0-content\" aria-expanded=\"false\">\n                    <i class=\"accordion-icon\"><svg  class=\"tf_fa tf-ti-plus\" aria-hidden=\"true\"><use href=\"#tf-ti-plus\"><\/use><\/svg><\/i>                    <i class=\"accordion-active-icon tf_hide\"><svg  class=\"tf_fa tf-ti-minus\" aria-hidden=\"true\"><use href=\"#tf-ti-minus\"><\/use><\/svg><\/i>                    <span class=\"accordion-title-wrap\">Optimizing Product Feature Offering through Connected Vehicles Analytics, Intelligent Sampling, and Generative AI <\/span>                <\/a>\n            <\/div><!-- .accordion-title -->\n            <div id=\"acc-fojv99-0-content\" data-id=\"acc-fojv99-0\" aria-hidden=\"true\" class=\"accordion-content tf_hide tf_clearfix\">\n                        <div  data-lazy=\"1\" class=\"module_subrow themify_builder_sub_row tf_w col_align_top tb_col_count_1 tb_8x26971\">\n                <div  data-lazy=\"1\" class=\"module_column sub_column tb_qz3f971 first\">\n                    <!-- module text -->\n<div  class=\"module module-text tb_rx7t971   \" data-lazy=\"1\">\n        <div  class=\"tb_text_wrap\">\n        <p><strong><span class=\"TextRun SCXW214232028 BCX8\" lang=\"EN-US\" xml:lang=\"EN-US\" data-contrast=\"none\"><span class=\"NormalTextRun SCXW214232028 BCX8\" data-ccp-parastyle=\"x_xmsonormal\" data-ccp-parastyle-defn=\"{&quot;ObjectId&quot;:&quot;5b4e8554-4129-4dfb-b302-00deccf256c7|45&quot;,&quot;ClassId&quot;:1073872969,&quot;Properties&quot;:[201342446,&quot;1&quot;,201342447,&quot;5&quot;,201342448,&quot;1&quot;,201342449,&quot;1&quot;,469777841,&quot;Times New Roman&quot;,469777842,&quot;Times New Roman&quot;,469777843,&quot;Times New Roman&quot;,469777844,&quot;Times New Roman&quot;,201341986,&quot;1&quot;,469769226,&quot;Times New Roman&quot;,268442635,&quot;24&quot;,469775450,&quot;x_xmsonormal&quot;,201340122,&quot;2&quot;,134233614,&quot;true&quot;,469778129,&quot;xxmsonormal&quot;,335572020,&quot;1&quot;,335559740,&quot;240&quot;,201341983,&quot;0&quot;,134233118,&quot;true&quot;,134233117,&quot;true&quot;,469778324,&quot;Normal&quot;]}\">Authors:<\/span><\/span><\/strong><span class=\"TextRun SCXW214232028 BCX8\" lang=\"EN-US\" xml:lang=\"EN-US\" data-contrast=\"none\"><span class=\"NormalTextRun SCXW214232028 BCX8\" data-ccp-parastyle=\"x_xmsonormal\"> Batchu, Vamsee Krishna<\/span><span class=\"NormalTextRun SCXW214232028 BCX8\" data-ccp-parastyle=\"x_xmsonormal\">;<\/span> <span class=\"NormalTextRun SpellingErrorV2Themed SCXW214232028 BCX8\" data-ccp-parastyle=\"x_xmsonormal\">Makke<\/span><span class=\"NormalTextRun SCXW214232028 BCX8\" data-ccp-parastyle=\"x_xmsonormal\">, Omar<\/span><span class=\"NormalTextRun SCXW214232028 BCX8\" data-ccp-parastyle=\"x_xmsonormal\">; <\/span><span class=\"NormalTextRun SpellingErrorV2Themed SCXW214232028 BCX8\" data-ccp-parastyle=\"x_xmsonormal\">Gusikhin<\/span><span class=\"NormalTextRun SCXW214232028 BCX8\" data-ccp-parastyle=\"x_xmsonormal\">, Oleg<\/span><span class=\"NormalTextRun SCXW214232028 BCX8\" data-ccp-parastyle=\"x_xmsonormal\">; <\/span><span class=\"NormalTextRun SpellingErrorV2Themed SCXW214232028 BCX8\" data-ccp-parastyle=\"x_xmsonormal\">Krivtsov<\/span><span class=\"NormalTextRun SCXW214232028 BCX8\" data-ccp-parastyle=\"x_xmsonormal\">, <\/span><span class=\"NormalTextRun SCXW214232028 BCX8\" data-ccp-parastyle=\"x_xmsonormal\">Vasiliy<\/span><span class=\"NormalTextRun SCXW214232028 BCX8\" data-ccp-parastyle=\"x_xmsonormal\">;\u00a0<\/span> <span class=\"NormalTextRun SpellingErrorV2Themed SCXW214232028 BCX8\" data-ccp-parastyle=\"x_xmsonormal\">Klampfl<\/span><span class=\"NormalTextRun SCXW214232028 BCX8\" data-ccp-parastyle=\"x_xmsonormal\">, <\/span><span class=\"NormalTextRun ContextualSpellingAndGrammarErrorV2Themed SCXW214232028 BCX8\" data-ccp-parastyle=\"x_xmsonormal\">Erica<\/span><span class=\"NormalTextRun ContextualSpellingAndGrammarErrorV2Themed SCXW214232028 BCX8\" data-ccp-parastyle=\"x_xmsonormal\">,<\/span><span class=\"NormalTextRun ContextualSpellingAndGrammarErrorV2Themed SCXW214232028 BCX8\" data-ccp-parastyle=\"x_xmsonormal\">;<\/span> <span class=\"NormalTextRun SpellingErrorV2Themed SCXW214232028 BCX8\" data-ccp-parastyle=\"x_xmsonormal\">Svidenko<\/span><span class=\"NormalTextRun SCXW214232028 BCX8\" data-ccp-parastyle=\"x_xmsonormal\">, Vicky<\/span><span class=\"NormalTextRun SCXW214232028 BCX8\" data-ccp-parastyle=\"x_xmsonormal\">;<\/span> <span class=\"NormalTextRun SCXW214232028 BCX8\" data-ccp-parastyle=\"x_xmsonormal\">Cooper, Rodney<\/span><span class=\"NormalTextRun SCXW214232028 BCX8\" data-ccp-parastyle=\"x_xmsonormal\">;<\/span> <span class=\"NormalTextRun SpellingErrorV2Themed SCXW214232028 BCX8\" data-ccp-parastyle=\"x_xmsonormal\">Shayrin<\/span><span class=\"NormalTextRun SCXW214232028 BCX8\" data-ccp-parastyle=\"x_xmsonormal\">,<\/span><span class=\"NormalTextRun SCXW214232028 BCX8\" data-ccp-parastyle=\"x_xmsonormal\"> Mare K<\/span><span class=\"NormalTextRun SCXW214232028 BCX8\" data-ccp-parastyle=\"x_xmsonormal\">;<\/span><\/span><span class=\"EOP SCXW214232028 BCX8\" data-ccp-props=\"{&quot;134233117&quot;:false,&quot;134233118&quot;:false,&quot;201341983&quot;:0,&quot;335557856&quot;:16777215,&quot;335559685&quot;:360,&quot;335559738&quot;:0,&quot;335559739&quot;:0,&quot;335559740&quot;:240}\">\u00a0<\/span><\/p>\n<p><span class=\"TextRun SCXW26939695 BCX8\" lang=\"EN-US\" xml:lang=\"EN-US\" data-contrast=\"auto\"><span class=\"NormalTextRun SCXW26939695 BCX8\">Abstract:<\/span><\/span><span class=\"TextRun SCXW26939695 BCX8\" lang=\"EN-US\" xml:lang=\"EN-US\" data-contrast=\"auto\"> <span class=\"NormalTextRun SCXW26939695 BCX8\">Automakers continuously look for ways to reduce <\/span><span class=\"NormalTextRun ContextualSpellingAndGrammarErrorV2Themed SCXW26939695 BCX8\">cost<\/span><span class=\"NormalTextRun SCXW26939695 BCX8\">, while not <\/span><span class=\"NormalTextRun SCXW26939695 BCX8\">impacting<\/span><span class=\"NormalTextRun SCXW26939695 BCX8\"> customer satisfaction in the products. By <\/span><span class=\"NormalTextRun SCXW26939695 BCX8\">leveraging<\/span><span class=\"NormalTextRun SCXW26939695 BCX8\"> connected vehicle data, we can <\/span><span class=\"NormalTextRun SCXW26939695 BCX8\">determine<\/span><span class=\"NormalTextRun SCXW26939695 BCX8\"> actual use under a variety of conditions and provide insights for product development to help justify adding or removing features or redesigning for use. The accelerated growth and heterogeneity of the data, which <\/span><span class=\"NormalTextRun SCXW26939695 BCX8\">result<\/span><span class=\"NormalTextRun SCXW26939695 BCX8\"> from constantly releasing new vehicle models with new capabilities appealing to different audiences, pose significant challenges related to collecting statistically representative data samples for a proper study. To address these challenges, an intelligent sampling system has been developed, <\/span><span class=\"NormalTextRun SCXW26939695 BCX8\">utilizing<\/span><span class=\"NormalTextRun SCXW26939695 BCX8\"> multi-agent Large Language Models (LLM) to <\/span><span class=\"NormalTextRun SCXW26939695 BCX8\">optimize<\/span><span class=\"NormalTextRun SCXW26939695 BCX8\"> data collection and analysis. This system centralizes and modularizes information retrieval for domain knowledge, signals, and existing strata. One exemplary application of this system is <\/span><span class=\"NormalTextRun ContextualSpellingAndGrammarErrorV2Themed SCXW26939695 BCX8\">feature<\/span><span class=\"NormalTextRun SCXW26939695 BCX8\"> rationalization, as <\/span><span class=\"NormalTextRun SCXW26939695 BCX8\">demonstrated<\/span><span class=\"NormalTextRun SCXW26939695 BCX8\"> by the removal of the parallel park <\/span><span class=\"NormalTextRun SCXW26939695 BCX8\">assist<\/span><span class=\"NormalTextRun SCXW26939695 BCX8\">, which resulted in substantial cost savings for the company and was widely publicized by key media outlets.<\/span><\/span><\/p>    <\/div>\n<\/div>\n<!-- \/module text -->        <\/div>\n                    <\/div>\n                    <\/div><!-- .accordion-content -->\n        <\/li>\n        <\/ul>\n\n<\/div><!-- \/module accordion --><!-- module accordion -->\n<div  class=\"module module-accordion tb_730448 \" data-behavior=\"toggle\" data-lazy=\"1\">\n    \n    <ul class=\"ui module-accordion   tb_default_color\">\n            <li>\n            <div class=\"accordion-title tf_rel\">\n                <a href=\"#acc-730448-0\" class=\"tb_title_accordion\" aria-controls=\"acc-730448-0-content\" aria-expanded=\"false\">\n                    <i class=\"accordion-icon\"><svg  class=\"tf_fa tf-ti-plus\" aria-hidden=\"true\"><use href=\"#tf-ti-plus\"><\/use><\/svg><\/i>                    <i class=\"accordion-active-icon tf_hide\"><svg  class=\"tf_fa tf-ti-minus\" aria-hidden=\"true\"><use href=\"#tf-ti-minus\"><\/use><\/svg><\/i>                    <span class=\"accordion-title-wrap\">Optimizing the Path Towards Plastic-Free Oceans<\/span>                <\/a>\n            <\/div><!-- .accordion-title -->\n            <div id=\"acc-730448-0-content\" data-id=\"acc-730448-0\" aria-hidden=\"true\" class=\"accordion-content tf_hide tf_clearfix\">\n                        <div  data-lazy=\"1\" class=\"module_subrow themify_builder_sub_row tf_w col_align_top tb_col_count_1 tb_pgw1971\">\n                <div  data-lazy=\"1\" class=\"module_column sub_column tb_ydxb971 first\">\n                    <!-- module text -->\n<div  class=\"module module-text tb_5yp2971   \" data-lazy=\"1\">\n        <div  class=\"tb_text_wrap\">\n        <p><strong><span class=\"TextRun SCXW26213306 BCX8\" lang=\"EN-US\" xml:lang=\"EN-US\" data-contrast=\"auto\"><span class=\"NormalTextRun SCXW26213306 BCX8\">Authors:<\/span><\/span><\/strong><span class=\"TextRun SCXW26213306 BCX8\" lang=\"EN-US\" xml:lang=\"EN-US\" data-contrast=\"auto\"> <span class=\"NormalTextRun SCXW26213306 BCX8\">Jean <\/span><span class=\"NormalTextRun SpellingErrorV2Themed SCXW26213306 BCX8\">Pauphilet<\/span><span class=\"NormalTextRun SCXW26213306 BCX8\"> &amp; <\/span><span class=\"NormalTextRun SpellingErrorV2Themed SCXW26213306 BCX8\">Baizhi<\/span><span class=\"NormalTextRun SCXW26213306 BCX8\"> Song<\/span><span class=\"NormalTextRun SCXW26213306 BCX8\"> (<\/span><span class=\"NormalTextRun SCXW26213306 BCX8\">London Business School<\/span><span class=\"NormalTextRun SCXW26213306 BCX8\">), <\/span><span class=\"NormalTextRun SCXW26213306 BCX8\">Yannick Pham &amp; Bruno Sainte<\/span><span class=\"NormalTextRun SCXW26213306 BCX8\">&#8211;<\/span><span class=\"NormalTextRun SCXW26213306 BCX8\">Rose<\/span><span class=\"NormalTextRun SCXW26213306 BCX8\"> (<\/span><span class=\"NormalTextRun SCXW26213306 BCX8\">The Ocean Cleanup<\/span><span class=\"NormalTextRun SCXW26213306 BCX8\">), <\/span><span class=\"NormalTextRun SCXW26213306 BCX8\">Dick den Hertog<\/span><span class=\"NormalTextRun SCXW26213306 BCX8\"> (<\/span><span class=\"NormalTextRun SCXW26213306 BCX8\">University of Amsterdam<\/span><span class=\"NormalTextRun SCXW26213306 BCX8\">)<\/span><\/span><span class=\"EOP SCXW26213306 BCX8\" data-ccp-props=\"{&quot;201341983&quot;:0,&quot;335559685&quot;:360,&quot;335559739&quot;:0,&quot;335559740&quot;:240}\">\u00a0<\/span><\/p>\n<p><span class=\"TextRun SCXW58163530 BCX8\" lang=\"EN-US\" xml:lang=\"EN-US\" data-contrast=\"auto\"><span class=\"NormalTextRun SCXW58163530 BCX8\">Increasing ocean plastic pollution is irreversibly harming ecosystems and human economic activities. We developed advanced predictive and prescriptive analytics methods to help The Ocean Cleanup to remove the plastic from the oceans faster. Specifically, we <\/span><span class=\"NormalTextRun SCXW58163530 BCX8\">optimize<\/span><span class=\"NormalTextRun SCXW58163530 BCX8\"> the route of their plastic collection system in the ocean to maximize the quantity of plastic collected over time. For the input for the optimization method, we developed a suite <\/span><span class=\"NormalTextRun SCXW58163530 BCX8\">of models to estimate the dispersion and density of plastic in the Great Pacific Garbage Patch, by using hindcast and forecast models of ocean currents, waves, and wind. The dispersion model uses fluid dynamics to simulate the transport of floating plastics in a time-dependent manner and is calibrated using empirical data. We formulate the problem as <\/span><span class=\"NormalTextRun ContextualSpellingAndGrammarErrorV2Themed SCXW58163530 BCX8\">a<\/span><span class=\"NormalTextRun SCXW58163530 BCX8\"> longest path problem in a well-structured graph. However, because collection directly <\/span><span class=\"NormalTextRun SCXW58163530 BCX8\">impacts<\/span><span class=\"NormalTextRun SCXW58163530 BCX8\"> future plastic density, the corresponding edge lengths are nonlinear polynomials. After analyzing the structural properties of the edge lengths, we propose a search-and-bound method, which <\/span><span class=\"NormalTextRun SCXW58163530 BCX8\">leverages<\/span><span class=\"NormalTextRun SCXW58163530 BCX8\"> a relaxation of the problem solvable via dynamic programming and clustering, to efficiently find high-quality solutions (within 6% <\/span><span class=\"NormalTextRun SCXW58163530 BCX8\">optimal<\/span><span class=\"NormalTextRun SCXW58163530 BCX8\"> in practice) and develop a tailored branch-and-bound strategy to solve it to provable optimality. On one year of ocean data, our optimization-based routing approach increases the quantity of plastic collected by more than 60% compared with the current routing strategy, hence speeding up the progress toward plastic-free oceans. Recently, the CEO of The Ocean Cleanup estimated that this innovative analytics <\/span><span class=\"NormalTextRun SCXW58163530 BCX8\">methodology<\/span><span class=\"NormalTextRun SCXW58163530 BCX8\"> halves the time needed to clean up the ocean to 5 <\/span><span class=\"NormalTextRun ContextualSpellingAndGrammarErrorV2Themed SCXW58163530 BCX8\">years, and<\/span><span class=\"NormalTextRun SCXW58163530 BCX8\"> nearly <\/span><span class=\"NormalTextRun ContextualSpellingAndGrammarErrorV2Themed SCXW58163530 BCX8\">halves<\/span><span class=\"NormalTextRun SCXW58163530 BCX8\"> the costs to <\/span><span class=\"NormalTextRun SCXW58163530 BCX8\">4 billion dollars<\/span><span class=\"NormalTextRun SCXW58163530 BCX8\">.\u00a0\u00a0<\/span><\/span><span class=\"EOP SCXW58163530 BCX8\" data-ccp-props=\"{&quot;335559685&quot;:450,&quot;335559731&quot;:360}\">\u00a0<\/span><\/p>    <\/div>\n<\/div>\n<!-- \/module text -->        <\/div>\n                    <\/div>\n                    <\/div><!-- .accordion-content -->\n        <\/li>\n        <\/ul>\n\n<\/div><!-- \/module accordion --><!-- module accordion -->\n<div  class=\"module module-accordion tb_w8mh511 \" data-behavior=\"toggle\" data-lazy=\"1\">\n    \n    <ul class=\"ui module-accordion   tb_default_color\">\n            <li>\n            <div class=\"accordion-title tf_rel\">\n                <a href=\"#acc-w8mh511-0\" class=\"tb_title_accordion\" aria-controls=\"acc-w8mh511-0-content\" aria-expanded=\"false\">\n                    <i class=\"accordion-icon\"><svg  class=\"tf_fa tf-ti-plus\" aria-hidden=\"true\"><use href=\"#tf-ti-plus\"><\/use><\/svg><\/i>                    <i class=\"accordion-active-icon tf_hide\"><svg  class=\"tf_fa tf-ti-minus\" aria-hidden=\"true\"><use href=\"#tf-ti-minus\"><\/use><\/svg><\/i>                    <span class=\"accordion-title-wrap\">Innovating Manufacturing: The Autonomous Factory with Digital Twin and Reinforcement Learning for Intelligent Operations and Efficiency <\/span>                <\/a>\n            <\/div><!-- .accordion-title -->\n            <div id=\"acc-w8mh511-0-content\" data-id=\"acc-w8mh511-0\" aria-hidden=\"true\" class=\"accordion-content tf_hide tf_clearfix\">\n                        <div  data-lazy=\"1\" class=\"module_subrow themify_builder_sub_row tf_w col_align_top tb_col_count_1 tb_82ky971\">\n                <div  data-lazy=\"1\" class=\"module_column sub_column tb_5cze971 first\">\n                    <!-- module text -->\n<div  class=\"module module-text tb_ehnu971   \" data-lazy=\"1\">\n        <div  class=\"tb_text_wrap\">\n        <p><strong><span class=\"TextRun SCXW164451823 BCX8\" lang=\"EN-US\" xml:lang=\"EN-US\" data-contrast=\"none\"><span class=\"NormalTextRun SCXW164451823 BCX8\">Authors:<\/span><\/span><\/strong> <span class=\"TextRun SCXW164451823 BCX8\" lang=\"EN-US\" xml:lang=\"EN-US\" data-contrast=\"none\"><span class=\"NormalTextRun SCXW164451823 BCX8\">Young Jae Jang (Korea Advanced Institute of Science and Technology \u2013 KAIST)<\/span> <span class=\"NormalTextRun SpellingErrorV2Themed SCXW164451823 BCX8\">Ilhoe<\/span><span class=\"NormalTextRun SCXW164451823 BCX8\"> Hwang and Seol Hwang (DAIM Research)<\/span><span class=\"NormalTextRun SCXW164451823 BCX8\">, <\/span><span class=\"NormalTextRun SpellingErrorV2Themed SCXW164451823 BCX8\">Sunhee<\/span><span class=\"NormalTextRun SCXW164451823 BCX8\"> Bang (Samsung SDI)<\/span><\/span><span class=\"EOP SCXW164451823 BCX8\" data-ccp-props=\"{&quot;201341983&quot;:0,&quot;335559685&quot;:360,&quot;335559739&quot;:0,&quot;335559740&quot;:240}\">\u00a0<\/span><\/p>\n<p><span class=\"NormalTextRun SCXW43265675 BCX8\">The &#8220;Autonomous Factory&#8221; (AF) concept revolutionizes factory operations by enabling systems to independently interpret environmental states and make decisions, much like autonomous vehicles. Developed collaboratively by Samsung SDI, KAIST, and DAIM Research, AF integrates reinforcement learning (RL) and Digital Twin technologies to address challenges in EV battery manufacturing, such as unpredictable events and dynamic production needs. The system resolves inefficiencies like deadlock issues in Autonomous Mobile Robot (AMR) systems, <\/span><span class=\"NormalTextRun SCXW43265675 BCX8\">eliminates<\/span><span class=\"NormalTextRun SCXW43265675 BCX8\"> the need for manual interventions, and improves operational efficiency. The AF system increased robot capacity by 21% and reduced the required fleet size from 500 to 400 robots, saving $8 million in robot investment costs. Additionally, the system reduces reliance on human operators, saving a minimum of $1 million annually in labor costs. This innovation sets a new standard in AI-driven manufacturing.<\/span><\/p>    <\/div>\n<\/div>\n<!-- \/module text -->        <\/div>\n                    <\/div>\n                    <\/div><!-- .accordion-content -->\n        <\/li>\n        <\/ul>\n\n<\/div><!-- \/module accordion -->        <\/div>\n                        <\/div>\n        <\/div>\n        <\/div>\n<!--\/themify_builder_content-->","protected":false},"excerpt":{"rendered":"<p>Track: INFORMS Prizes &#8211; IAAA Finalists<\/p>\n","protected":false},"author":1001140,"featured_media":0,"parent":0,"menu_order":0,"comment_status":"closed","ping_status":"closed","template":"","meta":{"_acf_changed":false,"content-type":"","footnotes":""},"class_list":["post-9735","page","type-page","status-publish","hentry","has-post-title","has-post-date","has-post-category","has-post-tag","has-post-comment","has-post-author",""],"acf":[],"yoast_head":"<!-- This site is optimized with the Yoast SEO Premium plugin v26.0 (Yoast SEO v26.0) - https:\/\/yoast.com\/wordpress\/plugins\/seo\/ -->\n<title>IAAA Finalists - 2025 INFORMS Analytics+ Conference<\/title>\n<meta name=\"description\" content=\"The IAAA Finalists will present at the 2025 INFORMS Analytics+ Conference.\" \/>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/meetings.informs.org\/wordpress\/analytics2025\/iaaa\/\" \/>\n<meta property=\"og:locale\" content=\"en_US\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"IAAA Finalists\" \/>\n<meta property=\"og:description\" content=\"The IAAA Finalists will present at the 2025 INFORMS Analytics+ Conference.\" \/>\n<meta property=\"og:url\" content=\"https:\/\/meetings.informs.org\/wordpress\/analytics2025\/iaaa\/\" \/>\n<meta property=\"og:site_name\" content=\"2025 INFORMS Analytics+ Conference\" \/>\n<meta property=\"article:publisher\" content=\"https:\/\/www.facebook.com\/INFORMSpage\/\" \/>\n<meta property=\"article:modified_time\" content=\"2026-03-05T14:27:51+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/meetings.informs.org\/wordpress\/analytics2025\/files\/2024\/10\/analyticsplus-logo.jpg\" \/>\n\t<meta property=\"og:image:width\" content=\"400\" \/>\n\t<meta property=\"og:image:height\" content=\"400\" \/>\n\t<meta property=\"og:image:type\" content=\"image\/jpeg\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:site\" content=\"@INFORMS\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"WebPage\",\"@id\":\"https:\/\/meetings.informs.org\/wordpress\/analytics2025\/iaaa\/\",\"url\":\"https:\/\/meetings.informs.org\/wordpress\/analytics2025\/iaaa\/\",\"name\":\"IAAA Finalists - 2025 INFORMS Analytics+ Conference\",\"isPartOf\":{\"@id\":\"https:\/\/meetings.informs.org\/wordpress\/analytics2025\/#website\"},\"datePublished\":\"2025-03-21T16:09:10+00:00\",\"dateModified\":\"2026-03-05T14:27:51+00:00\",\"description\":\"The IAAA Finalists will present at the 2025 INFORMS Analytics+ Conference.\",\"breadcrumb\":{\"@id\":\"https:\/\/meetings.informs.org\/wordpress\/analytics2025\/iaaa\/#breadcrumb\"},\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/meetings.informs.org\/wordpress\/analytics2025\/iaaa\/\"]}]},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\/\/meetings.informs.org\/wordpress\/analytics2025\/iaaa\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\/\/meetings.informs.org\/wordpress\/analytics2025\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"IAAA Finalists\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\/\/meetings.informs.org\/wordpress\/analytics2025\/#website\",\"url\":\"https:\/\/meetings.informs.org\/wordpress\/analytics2025\/\",\"name\":\"2025 INFORMS Analytics+ Conference\",\"description\":\"April 6-8, 2025\",\"publisher\":{\"@id\":\"https:\/\/meetings.informs.org\/wordpress\/analytics2025\/#organization\"},\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\/\/meetings.informs.org\/wordpress\/analytics2025\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"en-US\"},{\"@type\":\"Organization\",\"@id\":\"https:\/\/meetings.informs.org\/wordpress\/analytics2025\/#organization\",\"name\":\"INFORMS\",\"url\":\"https:\/\/meetings.informs.org\/wordpress\/analytics2025\/\",\"logo\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\/\/meetings.informs.org\/wordpress\/analytics2025\/#\/schema\/logo\/image\/\",\"url\":\"https:\/\/meetings.informs.org\/wordpress\/analytics2025\/files\/2022\/03\/INFORMS_Logo_full_color.jpg\",\"contentUrl\":\"https:\/\/meetings.informs.org\/wordpress\/analytics2025\/files\/2022\/03\/INFORMS_Logo_full_color.jpg\",\"width\":300,\"height\":72,\"caption\":\"INFORMS\"},\"image\":{\"@id\":\"https:\/\/meetings.informs.org\/wordpress\/analytics2025\/#\/schema\/logo\/image\/\"},\"sameAs\":[\"https:\/\/www.facebook.com\/INFORMSpage\/\",\"https:\/\/x.com\/INFORMS\",\"http:\/\/instagram.com\/informs_orms\",\"https:\/\/www.linkedin.com\/groups\/82644\/profile\",\"https:\/\/www.pinterest.com\/informs\/_saved\/\",\"https:\/\/www.youtube.com\/user\/INFORMSonline\",\"https:\/\/en.wikipedia.org\/wiki\/Institute_for_Operations_Research_and_the_Management_Sciences\"]}]}<\/script>\n<!-- \/ Yoast SEO Premium plugin. -->","yoast_head_json":{"title":"IAAA Finalists - 2025 INFORMS Analytics+ Conference","description":"The IAAA Finalists will present at the 2025 INFORMS Analytics+ Conference.","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/meetings.informs.org\/wordpress\/analytics2025\/iaaa\/","og_locale":"en_US","og_type":"article","og_title":"IAAA Finalists","og_description":"The IAAA Finalists will present at the 2025 INFORMS Analytics+ Conference.","og_url":"https:\/\/meetings.informs.org\/wordpress\/analytics2025\/iaaa\/","og_site_name":"2025 INFORMS Analytics+ Conference","article_publisher":"https:\/\/www.facebook.com\/INFORMSpage\/","article_modified_time":"2026-03-05T14:27:51+00:00","og_image":[{"width":400,"height":400,"url":"https:\/\/meetings.informs.org\/wordpress\/analytics2025\/files\/2024\/10\/analyticsplus-logo.jpg","type":"image\/jpeg"}],"twitter_card":"summary_large_image","twitter_site":"@INFORMS","schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"WebPage","@id":"https:\/\/meetings.informs.org\/wordpress\/analytics2025\/iaaa\/","url":"https:\/\/meetings.informs.org\/wordpress\/analytics2025\/iaaa\/","name":"IAAA Finalists - 2025 INFORMS Analytics+ Conference","isPartOf":{"@id":"https:\/\/meetings.informs.org\/wordpress\/analytics2025\/#website"},"datePublished":"2025-03-21T16:09:10+00:00","dateModified":"2026-03-05T14:27:51+00:00","description":"The IAAA Finalists will present at the 2025 INFORMS Analytics+ Conference.","breadcrumb":{"@id":"https:\/\/meetings.informs.org\/wordpress\/analytics2025\/iaaa\/#breadcrumb"},"inLanguage":"en-US","potentialAction":[{"@type":"ReadAction","target":["https:\/\/meetings.informs.org\/wordpress\/analytics2025\/iaaa\/"]}]},{"@type":"BreadcrumbList","@id":"https:\/\/meetings.informs.org\/wordpress\/analytics2025\/iaaa\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/meetings.informs.org\/wordpress\/analytics2025\/"},{"@type":"ListItem","position":2,"name":"IAAA Finalists"}]},{"@type":"WebSite","@id":"https:\/\/meetings.informs.org\/wordpress\/analytics2025\/#website","url":"https:\/\/meetings.informs.org\/wordpress\/analytics2025\/","name":"2025 INFORMS Analytics+ Conference","description":"April 6-8, 2025","publisher":{"@id":"https:\/\/meetings.informs.org\/wordpress\/analytics2025\/#organization"},"potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/meetings.informs.org\/wordpress\/analytics2025\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"en-US"},{"@type":"Organization","@id":"https:\/\/meetings.informs.org\/wordpress\/analytics2025\/#organization","name":"INFORMS","url":"https:\/\/meetings.informs.org\/wordpress\/analytics2025\/","logo":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/meetings.informs.org\/wordpress\/analytics2025\/#\/schema\/logo\/image\/","url":"https:\/\/meetings.informs.org\/wordpress\/analytics2025\/files\/2022\/03\/INFORMS_Logo_full_color.jpg","contentUrl":"https:\/\/meetings.informs.org\/wordpress\/analytics2025\/files\/2022\/03\/INFORMS_Logo_full_color.jpg","width":300,"height":72,"caption":"INFORMS"},"image":{"@id":"https:\/\/meetings.informs.org\/wordpress\/analytics2025\/#\/schema\/logo\/image\/"},"sameAs":["https:\/\/www.facebook.com\/INFORMSpage\/","https:\/\/x.com\/INFORMS","http:\/\/instagram.com\/informs_orms","https:\/\/www.linkedin.com\/groups\/82644\/profile","https:\/\/www.pinterest.com\/informs\/_saved\/","https:\/\/www.youtube.com\/user\/INFORMSonline","https:\/\/en.wikipedia.org\/wiki\/Institute_for_Operations_Research_and_the_Management_Sciences"]}]}},"builder_content":"<h2>Track: INFORMS Prizes - IAAA Finalists<\/h2>\n<p>The IAAA Finalists will present at the conference on Tuesday, April 8 in the INFORMS Prizes Track from 9:10am \u2013 10:00am EST. Judges will then review the rank and announce the winner at the INFORMS Analytics Society luncheon.<\/p>\n<ul><li><h4>Spot Market Pricing on Amazon Freight<\/h4><p><b>Authors:<\/b> Adam N. Elmachtoub , Goutam Kumar, Roger Lederman\u00a0<\/p> <p><b>Affiliation:<\/b> Amazon.\u00a0<\/p> <p>Amazon Freight (AF) is a business unit at Amazon that moves freight for other\u00a0<\/p> <p>shippers using Amazon\u2019s internal middle mile network. For this business, external shippers\u00a0<\/p> <p>use a website to request a price quote for an origin-destination pair, get prices instantly,\u00a0<\/p> <p>and decide which option to book. AF provides prices that varies across lead time options.\u00a0<\/p> <p>The execution costs and demand can vary widely across different quotes due to internal cost\u00a0<\/p> <p>structure and overall freight market. The goal for AF is to set prices for each quote in order\u00a0<\/p> <p>to maximize a business objective. Our solution utilizes a combination of parameteric and\u00a0<\/p> <p>non-parametric modeling with price optimization. We first use quote features to segment\u00a0<\/p> <p>the market into disjoint groups, employing the Market Segmentation Tree (MST) algorithm\u00a0<\/p> <p>which creates a binary tree based on differences in choice behavior. Within each leaf, we fit\u00a0<\/p> <p>a reference-price-effects Multinomial (MNL) choice model that is amenable to fast pricing\u00a0<\/p> <p>heuristics. We conducted live A\/B experiments that shows our new framework significantly\u00a0<\/p> <p>improves profit, revenue, and accuracy. The model has been in production for about a year.\u00a0<\/p><\/li><\/ul>\n<p><b>Authors:<\/b> Adam N. Elmachtoub , Goutam Kumar, Roger Lederman\u00a0<\/p> <p><b>Affiliation:<\/b> Amazon.\u00a0<\/p> <p>Amazon Freight (AF) is a business unit at Amazon that moves freight for other\u00a0<\/p> <p>shippers using Amazon\u2019s internal middle mile network. For this business, external shippers\u00a0<\/p> <p>use a website to request a price quote for an origin-destination pair, get prices instantly,\u00a0<\/p> <p>and decide which option to book. AF provides prices that varies across lead time options.\u00a0<\/p> <p>The execution costs and demand can vary widely across different quotes due to internal cost\u00a0<\/p> <p>structure and overall freight market. The goal for AF is to set prices for each quote in order\u00a0<\/p> <p>to maximize a business objective. Our solution utilizes a combination of parameteric and\u00a0<\/p> <p>non-parametric modeling with price optimization. We first use quote features to segment\u00a0<\/p> <p>the market into disjoint groups, employing the Market Segmentation Tree (MST) algorithm\u00a0<\/p> <p>which creates a binary tree based on differences in choice behavior. Within each leaf, we fit\u00a0<\/p> <p>a reference-price-effects Multinomial (MNL) choice model that is amenable to fast pricing\u00a0<\/p> <p>heuristics. We conducted live A\/B experiments that shows our new framework significantly\u00a0<\/p> <p>improves profit, revenue, and accuracy. The model has been in production for about a year.\u00a0<\/p>\n<ul><li><h4>Semantically Grounded Analytics Engine for Helping Patients Find Doctors and Locations in a Large Healthcare Organization <\/h4><p><b>Authors:<\/b> Mayank Kejriwal*, Hamid Haidarian**, Min-Hsueh Chiu*, Andy Xiang**, Deep Shrestha**, Faizan Javed**\u00a0<\/p> <p><b>Affiliations:<\/b>\u202f*University of Southern California; **Kaiser Permanente Digital\u00a0<\/p> <p>Efficiently finding doctors and locations (FDL) is an important search problem for patients in the healthcare domain, for which traditional analytical and information retrieval approaches tend to be sub-optimal. Legacy FDL systems, based usually on keyword and simple pattern matching, lack\u202fsemantic\u202ffoundations, preventing them from understanding users\u2019 underlying intent. Through innovations in all three prongs of analytics (descriptive, predictive, and prescriptive), we present a solution to this problem that uses a rich combination of a healthcare knowledge graph and concept-rich ontology. The solution has been implemented as a real-time system for FDL in Kaiser Permanente, a large healthcare organization with 12 million members, and offers significant performance benefits while maintaining practical needs of data security and privacy, scalability, cost-effectiveness, and backward compatibility with existing search infrastructure<\/p><\/li><\/ul>\n<p><b>Authors:<\/b> Mayank Kejriwal*, Hamid Haidarian**, Min-Hsueh Chiu*, Andy Xiang**, Deep Shrestha**, Faizan Javed**\u00a0<\/p> <p><b>Affiliations:<\/b>\u202f*University of Southern California; **Kaiser Permanente Digital\u00a0<\/p> <p>Efficiently finding doctors and locations (FDL) is an important search problem for patients in the healthcare domain, for which traditional analytical and information retrieval approaches tend to be sub-optimal. Legacy FDL systems, based usually on keyword and simple pattern matching, lack\u202fsemantic\u202ffoundations, preventing them from understanding users\u2019 underlying intent. Through innovations in all three prongs of analytics (descriptive, predictive, and prescriptive), we present a solution to this problem that uses a rich combination of a healthcare knowledge graph and concept-rich ontology. The solution has been implemented as a real-time system for FDL in Kaiser Permanente, a large healthcare organization with 12 million members, and offers significant performance benefits while maintaining practical needs of data security and privacy, scalability, cost-effectiveness, and backward compatibility with existing search infrastructure<\/p>\n<ul><li><h4>Leveraging AI for AI , Auxiliary Models to sustain a customer facing chatbot<\/h4><p><strong>Authors:<\/strong> Fabio Dutra Sarti, Safa Shamout, Mohan Mulakkamparambil, Andres Villegas Ceballos,\u00a0 Antonio Henrique, Rafal Orlowski\u00a0<\/p> <p>This presentation explores the evolution of Scotiabank's customer-facing chatbot, from its initial launch and manual maintenance to the automation of key sustainment tasks using machine learning and AI. Scotiabank's \"AI for AI\" system, leveraging five proprietary AI models, significantly enhances its award-winning customer support chatbot. This innovative approach automates manual processes, streamlines development, and enables continuous improvement, resulting in reduced workload, enhanced chatbot performance, and increased customer satisfaction. Specifically, it details how auxiliary AI efficiently manages and scales a large-scale chatbot solution, saving thousands of manual work hours. Furthermore, the presentation will highlight the role of data analytics \u2013 through insightful reporting and A\/B testing \u2013 in optimizing customer experience and outcomes. Finally, it will address how Scotiabank has successfully implemented generative AI within a regulated environment.\u00a0<\/p><\/li><\/ul>\n<p><strong>Authors:<\/strong> Fabio Dutra Sarti, Safa Shamout, Mohan Mulakkamparambil, Andres Villegas Ceballos,\u00a0 Antonio Henrique, Rafal Orlowski\u00a0<\/p> <p>This presentation explores the evolution of Scotiabank's customer-facing chatbot, from its initial launch and manual maintenance to the automation of key sustainment tasks using machine learning and AI. Scotiabank's \"AI for AI\" system, leveraging five proprietary AI models, significantly enhances its award-winning customer support chatbot. This innovative approach automates manual processes, streamlines development, and enables continuous improvement, resulting in reduced workload, enhanced chatbot performance, and increased customer satisfaction. Specifically, it details how auxiliary AI efficiently manages and scales a large-scale chatbot solution, saving thousands of manual work hours. Furthermore, the presentation will highlight the role of data analytics \u2013 through insightful reporting and A\/B testing \u2013 in optimizing customer experience and outcomes. Finally, it will address how Scotiabank has successfully implemented generative AI within a regulated environment.\u00a0<\/p>\n<ul><li><h4>Optimizing Product Feature Offering through Connected Vehicles Analytics, Intelligent Sampling, and Generative AI <\/h4><p><strong>Authors:<\/strong> Batchu, Vamsee Krishna; Makke, Omar; Gusikhin, Oleg; Krivtsov, Vasiliy;\u00a0 Klampfl, Erica,; Svidenko, Vicky; Cooper, Rodney; Shayrin, Mare K;\u00a0<\/p> <p>Abstract: Automakers continuously look for ways to reduce cost, while not impacting customer satisfaction in the products. By leveraging connected vehicle data, we can determine actual use under a variety of conditions and provide insights for product development to help justify adding or removing features or redesigning for use. The accelerated growth and heterogeneity of the data, which result from constantly releasing new vehicle models with new capabilities appealing to different audiences, pose significant challenges related to collecting statistically representative data samples for a proper study. To address these challenges, an intelligent sampling system has been developed, utilizing multi-agent Large Language Models (LLM) to optimize data collection and analysis. This system centralizes and modularizes information retrieval for domain knowledge, signals, and existing strata. One exemplary application of this system is feature rationalization, as demonstrated by the removal of the parallel park assist, which resulted in substantial cost savings for the company and was widely publicized by key media outlets.<\/p><\/li><\/ul>\n<p><strong>Authors:<\/strong> Batchu, Vamsee Krishna; Makke, Omar; Gusikhin, Oleg; Krivtsov, Vasiliy;\u00a0 Klampfl, Erica,; Svidenko, Vicky; Cooper, Rodney; Shayrin, Mare K;\u00a0<\/p> <p>Abstract: Automakers continuously look for ways to reduce cost, while not impacting customer satisfaction in the products. By leveraging connected vehicle data, we can determine actual use under a variety of conditions and provide insights for product development to help justify adding or removing features or redesigning for use. The accelerated growth and heterogeneity of the data, which result from constantly releasing new vehicle models with new capabilities appealing to different audiences, pose significant challenges related to collecting statistically representative data samples for a proper study. To address these challenges, an intelligent sampling system has been developed, utilizing multi-agent Large Language Models (LLM) to optimize data collection and analysis. This system centralizes and modularizes information retrieval for domain knowledge, signals, and existing strata. One exemplary application of this system is feature rationalization, as demonstrated by the removal of the parallel park assist, which resulted in substantial cost savings for the company and was widely publicized by key media outlets.<\/p>\n<ul><li><h4>Optimizing the Path Towards Plastic-Free Oceans<\/h4><p><strong>Authors:<\/strong> Jean Pauphilet &amp; Baizhi Song (London Business School), Yannick Pham &amp; Bruno Sainte-Rose (The Ocean Cleanup), Dick den Hertog (University of Amsterdam)\u00a0<\/p> <p>Increasing ocean plastic pollution is irreversibly harming ecosystems and human economic activities. We developed advanced predictive and prescriptive analytics methods to help The Ocean Cleanup to remove the plastic from the oceans faster. Specifically, we optimize the route of their plastic collection system in the ocean to maximize the quantity of plastic collected over time. For the input for the optimization method, we developed a suite of models to estimate the dispersion and density of plastic in the Great Pacific Garbage Patch, by using hindcast and forecast models of ocean currents, waves, and wind. The dispersion model uses fluid dynamics to simulate the transport of floating plastics in a time-dependent manner and is calibrated using empirical data. We formulate the problem as a longest path problem in a well-structured graph. However, because collection directly impacts future plastic density, the corresponding edge lengths are nonlinear polynomials. After analyzing the structural properties of the edge lengths, we propose a search-and-bound method, which leverages a relaxation of the problem solvable via dynamic programming and clustering, to efficiently find high-quality solutions (within 6% optimal in practice) and develop a tailored branch-and-bound strategy to solve it to provable optimality. On one year of ocean data, our optimization-based routing approach increases the quantity of plastic collected by more than 60% compared with the current routing strategy, hence speeding up the progress toward plastic-free oceans. Recently, the CEO of The Ocean Cleanup estimated that this innovative analytics methodology halves the time needed to clean up the ocean to 5 years, and nearly halves the costs to 4 billion dollars.\u00a0\u00a0\u00a0<\/p><\/li><\/ul>\n<p><strong>Authors:<\/strong> Jean Pauphilet &amp; Baizhi Song (London Business School), Yannick Pham &amp; Bruno Sainte-Rose (The Ocean Cleanup), Dick den Hertog (University of Amsterdam)\u00a0<\/p> <p>Increasing ocean plastic pollution is irreversibly harming ecosystems and human economic activities. We developed advanced predictive and prescriptive analytics methods to help The Ocean Cleanup to remove the plastic from the oceans faster. Specifically, we optimize the route of their plastic collection system in the ocean to maximize the quantity of plastic collected over time. For the input for the optimization method, we developed a suite of models to estimate the dispersion and density of plastic in the Great Pacific Garbage Patch, by using hindcast and forecast models of ocean currents, waves, and wind. The dispersion model uses fluid dynamics to simulate the transport of floating plastics in a time-dependent manner and is calibrated using empirical data. We formulate the problem as a longest path problem in a well-structured graph. However, because collection directly impacts future plastic density, the corresponding edge lengths are nonlinear polynomials. After analyzing the structural properties of the edge lengths, we propose a search-and-bound method, which leverages a relaxation of the problem solvable via dynamic programming and clustering, to efficiently find high-quality solutions (within 6% optimal in practice) and develop a tailored branch-and-bound strategy to solve it to provable optimality. On one year of ocean data, our optimization-based routing approach increases the quantity of plastic collected by more than 60% compared with the current routing strategy, hence speeding up the progress toward plastic-free oceans. Recently, the CEO of The Ocean Cleanup estimated that this innovative analytics methodology halves the time needed to clean up the ocean to 5 years, and nearly halves the costs to 4 billion dollars.\u00a0\u00a0\u00a0<\/p>\n<ul><li><h4>Innovating Manufacturing: The Autonomous Factory with Digital Twin and Reinforcement Learning for Intelligent Operations and Efficiency <\/h4><p><strong>Authors:<\/strong> Young Jae Jang (Korea Advanced Institute of Science and Technology \u2013 KAIST) Ilhoe Hwang and Seol Hwang (DAIM Research), Sunhee Bang (Samsung SDI)\u00a0<\/p> <p>The \"Autonomous Factory\" (AF) concept revolutionizes factory operations by enabling systems to independently interpret environmental states and make decisions, much like autonomous vehicles. Developed collaboratively by Samsung SDI, KAIST, and DAIM Research, AF integrates reinforcement learning (RL) and Digital Twin technologies to address challenges in EV battery manufacturing, such as unpredictable events and dynamic production needs. The system resolves inefficiencies like deadlock issues in Autonomous Mobile Robot (AMR) systems, eliminates the need for manual interventions, and improves operational efficiency. The AF system increased robot capacity by 21% and reduced the required fleet size from 500 to 400 robots, saving $8 million in robot investment costs. Additionally, the system reduces reliance on human operators, saving a minimum of $1 million annually in labor costs. This innovation sets a new standard in AI-driven manufacturing.<\/p><\/li><\/ul>\n<p><strong>Authors:<\/strong> Young Jae Jang (Korea Advanced Institute of Science and Technology \u2013 KAIST) Ilhoe Hwang and Seol Hwang (DAIM Research), Sunhee Bang (Samsung SDI)\u00a0<\/p> <p>The \"Autonomous Factory\" (AF) concept revolutionizes factory operations by enabling systems to independently interpret environmental states and make decisions, much like autonomous vehicles. Developed collaboratively by Samsung SDI, KAIST, and DAIM Research, AF integrates reinforcement learning (RL) and Digital Twin technologies to address challenges in EV battery manufacturing, such as unpredictable events and dynamic production needs. The system resolves inefficiencies like deadlock issues in Autonomous Mobile Robot (AMR) systems, eliminates the need for manual interventions, and improves operational efficiency. The AF system increased robot capacity by 21% and reduced the required fleet size from 500 to 400 robots, saving $8 million in robot investment costs. Additionally, the system reduces reliance on human operators, saving a minimum of $1 million annually in labor costs. This innovation sets a new standard in AI-driven manufacturing.<\/p>","_links":{"self":[{"href":"https:\/\/meetings.informs.org\/wordpress\/analytics2025\/wp-json\/wp\/v2\/pages\/9735","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/meetings.informs.org\/wordpress\/analytics2025\/wp-json\/wp\/v2\/pages"}],"about":[{"href":"https:\/\/meetings.informs.org\/wordpress\/analytics2025\/wp-json\/wp\/v2\/types\/page"}],"author":[{"embeddable":true,"href":"https:\/\/meetings.informs.org\/wordpress\/analytics2025\/wp-json\/wp\/v2\/users\/1001140"}],"replies":[{"embeddable":true,"href":"https:\/\/meetings.informs.org\/wordpress\/analytics2025\/wp-json\/wp\/v2\/comments?post=9735"}],"version-history":[{"count":17,"href":"https:\/\/meetings.informs.org\/wordpress\/analytics2025\/wp-json\/wp\/v2\/pages\/9735\/revisions"}],"predecessor-version":[{"id":10118,"href":"https:\/\/meetings.informs.org\/wordpress\/analytics2025\/wp-json\/wp\/v2\/pages\/9735\/revisions\/10118"}],"wp:attachment":[{"href":"https:\/\/meetings.informs.org\/wordpress\/analytics2025\/wp-json\/wp\/v2\/media?parent=9735"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}