>> OK. OK, LET'S GO AHEAD AND GET STARTED. A COUPLE OF REMINDERS. WE HAVE A NEW TIME FOR THE SPEAKERS OVER THERE. IT WILL GIVE YOU 20 MINUTES. ABOUT 30 MINUTES TOTAL FOR EACH. ABOUT 10 MINUTES OF DISCUSSION. IF YOU NEED TO HAVE DISCUSSIONS, FEEL FREE TO GO OUTSIDE OR TALK IN THE HALLWAY. I WILL TRY TO KEEP THIS ON SCHEDULE SO EVERYONE CAN GET OUT. WE GET STARTED TODAY. FROM THE UNIVERSITY OF CALIFORNIA. >> HELLO. [INDISCERNIBLE] ADDITIONAL -- ADDITIONAL INFORMATION. BACKGROUND AND MOTIVATION. THEN WE REVEAL RECENT INFORMATION ON ROBUST UC AND COLUMN-AND CONSTRAIGHT. AND THE GENERATION METHOD. AN MATHEMATICAL FORMULATIONS AND RESULTS. AND THE ALGORITHM IMPROVEMENT. NEXT SLIDE. EVERYBODY KNOWS -- ONCE THE TRANSLATION -- TRANSMISSION LINES ARE THERE, YOU CAN NOW MOVE IT. THEN, SOME DYSFUNCTION. AT THAT MOMENT, WE ARE USING TRANSMISSION. [INDISCERNIBLE] RECENT INITIATIVE ON CALL OPTIMIZATION [INDISCERNIBLE] WE CAN ENSURE BETTER PERFORMANCE. DISPATCH COST CAN BE SAVED UP TO 25%. [INDISCERNIBLE] IF WE CAN INCORPORATE STRATEGY INTO THE OPERATION, [INDISCERNIBLE] THAT IS OUR TARGET. WE WANT TO COME TOGETHER . [INDISCERNIBLE] WHAT IS THE MOST INSTRUCTIVE CONTINGENCY. WE DO NOT KNOW. NOT ONLY IS IT UPGRADING, OR EXPANSION, THIS NEW LIFE HAS BENEFITS. [INDISCERNIBLE] HIGH-VOLTAGE IN THE SYSTEM. ROBUST UNIT COMMITMENT -- [INDISCERNIBLE] THIS PROJECT IS ONGOING. WE DID NOT FINISH. THANK YOU. ROBUST COMMENT A STRAIGHT -- AND CONSTRAINT. [INDISCERNIBLE] TUESDAY'S ROBUST UNIT COMMITMENT. THE FIRST ONE, WHICH IS [INDISCERNIBLE] IT IS CLOSED AND THEN WE BEGIN TO DO [INDISCERNIBLE] ALWAYS TRY TO PUSH YOUR COSTS AS HIGH AS POSSIBLE. WE HAVE THREE LEVELS. MINIMIZATION, MAXIMIZATION. TWO ALGORITHMS DEVELOPED, INCLUDING -- TWO ALGORITHMS, [INDISCERNIBLE] IT IS VERY SLOW. WE HAVE A NEW STRATEGY. THIS NEW ALGORITHM HAS TO DOES FEATURES. FIRST, MATHEMATICALLY, [INDISCERNIBLE] SECOND -- [INDISCERNIBLE] THE ALGORITHM HAS TO DOES PARTS. WE SAW THE SOLUTION. [INDISCERNIBLE] WE WANTED TO MAXIMIZE OR MINIMIZE [INDISCERNIBLE] THE ACTUAL QUESTION IS WHAT [INDISCERNIBLE] WE INTRODUCE NEW VARIABLES FOR EACH SCENARIO. [INDISCERNIBLE] IT IS VERY SMALL. VERY SMALL. THE ALGORITHM [INDISCERNIBLE] [INDISCERNIBLE] WE BRING IN THOSE ALGORITHMS -- IF YOU WANT TO KNOW THE DETAILS, YOU CAN CHECK THIS PAPER. WE ARE SHOWING THE COMPONENTS OF THE TWO ALGORITHMS. WE SHOW THE COMPOUND [INDISCERNIBLE] BA, I HIGHLIGHTED THESE, NATION TIMES. MORE THAN 20,000. IT IS HUGE. THE CRAZY FOR NET DEATH SCENARIO FOR US IS VERY SMALL. SMALL NUMBER. THAT IS WHY THE ALGORITHMS DO SO WELL. WE DO NOT WANT [INDISCERNIBLE] [INDISCERNIBLE] TIMER JUST FIVE HOURS IN ONE CASE. -- TIME REDUCED FIVE HOURS IN ONE CASE. [INDISCERNIBLE] YOU CAN TRY THESE ALGORITHMS. I AM HAPPY TO PROVIDE. HIGHLIGHTS. GENERAL -- MINIMIZATION, MAXIMIZATION, MINIMIZATION. A THREE LEVEL PROBLEM. THANK YOU. SO WE HAVE TWO FELL -- BINARY IS VAL -- BINARY VARIABLES. IT IS OUR ENEMY AND TRIES TO FORCE AN OUTAGE. [INDISCERNIBLE] THIS ONE IS BINARY. WE DO NOT HAVE DUALITY. [INDISCERNIBLE] RIGHT NOW, WE DO NOT HAVE A DEAL. [INDISCERNIBLE] IT IS AFFECTED BY TO DOES MARBLES. -- VARIABLES. WE WANT TO IDENTIFY [INDISCERNIBLE] IF WE CAN IDENTIFY THIS ONE, WE CAN DEVELOP A STRATEGY. WE CAN AVOID THIS. THIS HAS BEEN PRESENTED. PROVIDE EVIDENCE OF THIS PROBLEM WITH EMERGING COSTS. THAT AGAIN TO STOP WHAT WE DEVELOPED. [INDISCERNIBLE] THESE TWO VARIABLES, WE KEEP THEM, [INDISCERNIBLE] THIS ONE, OUTAGE, [INDISCERNIBLE] IDENTICAL TO THE [INDISCERNIBLE] WE TRIED. WE BENCHMARK OUR RESULTS. [INDISCERNIBLE] THE BENCHMARK I BELIEVE WILL HELP HERE. SO, THE TABLE IS [INDISCERNIBLE] ON AVERAGE, SOLVING THE PROBLEM 920 SECONDS. 17 SECONDS. HUGE. FURTHERMORE, THEY CANNOT VERIFY THE SOLUTION. [INDISCERNIBLE] IT DRASTICALLY OUTPERFORMS. IT IS A SMALL NUMBER. [INDISCERNIBLE] SO, NOW WE SHOWED -- WHEN K IS GREATER OR EQUAL TO FOUR, [INDISCERNIBLE] THE BENEFIT [INDISCERNIBLE] NOW WE SEE -- IS CALLING. [INDISCERNIBLE] [INDISCERNIBLE] THERE IS NO ALGORITHM AVAILABLE. [INDISCERNIBLE] REALLY HOT. WE CAN USE OUR ALGORITHM TO SOLVE THE IMPOUND. THAT IS THE WAY. [INDISCERNIBLE] THIS IDENTIFY THE WORST CASE. FOR THIS ALGORITHM, ENHANCEMENTS ARE CRITICAL. ONE EXAMPLE. SIMPLIFIED VERSION. [INDISCERNIBLE] IT WILL BE MORE, IT WILL BE STRONGER. SO WHICH LINE [INDISCERNIBLE] THIS IS ONE QUESTION. THE WORST CASE IS THREE RED LINE. [INDISCERNIBLE] THE YELLOW LINE REPRESENTING IN THE WORST CASE -- THE YELLOW LINES WILL BE SWITCHED OFF. HERE, AGAIN, OUR ALGORITHM. [INDISCERNIBLE] [INDISCERNIBLE] RHYTHM IMPROVEMENTS. -- ALGORITHM IMPROVEMENTS. [INDISCERNIBLE] INCLUDING THE ALGORITHM IS VERY CRITICAL. [INDISCERNIBLE] BELIEVE VISUALIZED A WORKPLACE ACCORDING TO OUR INTELLIGENCE. WE CAN CREATE A WORST-CASE. ANY SCENARIO, IT WILL CAUSE TROUBLE. [INDISCERNIBLE] MARGINAL SOLUTIONS AT THE SAME TIME. [INDISCERNIBLE] THEN WE ADJUST OUR OPTIMUM TOLERANCE. MORE TRICKS, WE WELCOME. THIS ALGORITHM -- YOU HAVE TROUBLE. [INDISCERNIBLE] BY EXAMPLE -- WE WANTED TO GENERATE A DATA SET FOR THE VULNERABILITY ANALYSIS. THIS SHOWS OUR BASIC METHOD. THIS COLUMN. HERE, IT IS UNBEARABLE. TOO MUCH. [INDISCERNIBLE] MAKES THE ALGORITHM BETTER. [INDISCERNIBLE] SOMETIMES, GETTING WORSE. MOST LIKELY, THEY ARE GETTING BETTER. IF WE TAKE AVERAGE, IN THE BEGINNING, IT IS ONE OF 800. IT WILL BE REDUCED TO 200. SO SOME INTELLIGENCE HERE COMES OUR -- IT IS FAR BETTER. [INDISCERNIBLE] [INDISCERNIBLE] QUESTIONS? I AM LOOKING FORWARD TO QUESTIONS. CHRIST THANK YOU VERY MUCH. -- >> THANK YOU VERY MUCH. [APPLAUSE] .