>>> GOOD MORNING. THANK YOU VERY MUCH FOR YOUR ATTENTION. MY NAME IS JOSHUA, A PHD STUDENT OF INDUSTRIAL ENGINEERING. TODAY, I WILL BE TALKING SOME ABOUT SOME OF THE WORK WE ARE DOING AS IT RELATES TO RESERVE REQUIREMENTS AND, IN PARTICULAR, HOW TO ENSURE WHEN WE REACTIVATE , WE CAN DO SO IN A WAY THAT DOES NOT VIOLATE [INDISCERNIBLE] BEFORE TALKING ABOUT RESERVES IN PARTICULAR, I WOULD LIKE TO SPEND A MOMENT TALKING ABOUT THE PREVIOUS ZONE ENERGY MARKET. IN THE EARLY TO MID-2000, TEXAS'S ENERGY MARKET ONLY MODELED ON THESE ZONES. ONE OF THE MAJOR ASSUMPTIONS THEY WERE MAKING IS THAT THE SECTION WITHIN THE ZONES WOULD BE NEGLIGIBLE. IN CASES WHEN IT WAS NOT, WHICH HAPPENS, THEY WOULD THEN ADJUST THE SOLUTION AFTER THE FACT TO GET SOMETHING THAT WAS TECHNICALLY FEASIBLE. WHICH THIS IS SHOWING, IN 2001, THEY SPENT A RELATIVELY SMALL ABOUT -- AMOUNT MITIGATING LOCAL CONGESTION. THAT ASSUMPTION SEEMS TO BE WORKING WELL. OVER TIME, IT BECAME A LESS AND LESS REASONABLE THING. THERE ARE A COUPLE OF TAKEAWAYS FROM THIS. THE ZONES DO NOT PERFECTLY CHARACTERIZED SUGGESTIONS. IN CASES WHERE THEY DO NOT, WE WOULD LIKE TO HAVE EFFECTIVE WAYS OF ADJUSTING THE SOLUTION SO WE DO NOT HAVE TO LIVE WITH THE CONSEQUENCES OF ONE OF THOSE FOUR PROCLAMATIONS. SOME OF THESE ARGUMENTS STILL APPLY. TODAY, I WILL AGAIN TALK ABOUT THE WORK WE'RE DOING. I WILL BE GIVING A HIGH-LEVEL OVERVIEW OF THREE APPROACHES WE ARE WORKING ON AND ON -- WILL BE PRESENTING THEM -- A RESERVE ZONE DETERMINATION WE ARE DOING OFF-LINE, IN ORDER TO BETTER ANTICIPATE WHAT THEY WILL BE, A CONGRATS -- CONGESTION RESERVE REQUIREMENT, AND AN AFTER-THE- FACT METHODOLOGY FOR SOLUTIONS IN ORDER TO PREP FOR INVALID ASSUMPTIONS. OUR MOTIVATION IS TO IMPROVE EXISTING RESERVE POLICIES, AS WELL AS TO CREATE BETTER POLITIES THAT CAN BETTER HANDLE. RESERVE REQUIREMENTS OR GRANTING REQUIREMENTS, AS WELL, A SIMILAR THING, THEY ENSURE THE CAPACITY THAT ALLOW OPERATORS TO RESPOND TO THESE SERVICES IN ORDER TO REACH OUR FREQUENCY. HOWEVER, THE RESERVE REQUIREMENTS IN ANOTHER THEMSELVES ONLY ENSURE CAPACITY, PHYSICAL AVAILABILITY OF THE RESERVE, BUT THAT DOES NOT NECESSARILY MEAN IT IS TENTATIVELY AVAILABLE. EVEN THOUGH IT EXISTS, IT MIGHT NOT BE AVAILABLE TO YOU. ENSURING THIS RESERVE IS DELIVERABLE IS NOT NECESSARILY A TRIPLE LEO -- TRIVIAL PROBLEM. IT WILL BE MORE DIFFICULT WHEN YOU HAVE HIGHER UNCERTAINTY OR HIGH PENETRATIONS OF WIND AND SOLAR. THE LITERATURE IS ACTUALLY PRETTY IN-DEPTH. USUALLY PROTECTING AGAINST -- OFTEN PROTECTING AGAINST A CERTAIN AMOUNT OF OUTAGES OR BALANCING THE VALUE, THE COST OF RESERVE WITH VALUE. ENSURING THE RESERVE IS DELIVERABLE IS OFTEN LOOKED AT AS A SEPARATE PROCESS WHERE YOU USE ZONES TO LOCATE RESERVES, OR YOU USE VARIOUS TRANSITION MARGINS TO ALLOW FLEXIBILITY TO DISPATCH RESERVE ONCE UNCERTAINTY REVEALS ITSELF. AN INTERESTING AREA IS THE INTERSECTION. THIS IS WHERE WE FIND A LOT OF THE THINGS WE ARE TALKING ABOUT. ROBUST OPTIMIZATION. THIS IS REALLY THE PREFERRED -- AREA TO OPERATE. WE CAN CO-OPTIMIZE THESE POSITIONS AS THEY RELATE TO THE RESERVE AND ENSURE RELIABILITY IN ANY MORE MANNER. THE CHALLENGE, HOWEVER, IS SOME OF THESE PROBLEMS CAN BE DIFFICULT TO SOLVE, SO ONE OF THE THINGS, A COMMON RECOMMENDATION THROUGHOUT THE PROGRAMMING IS THEY WILL SAY, WE WILL LIMIT THIS TO A LIMITED NUMBER OF SCENARIOS AND THEN USE DETERMINISTIC POLICIES TO HANDLE THE REMAINING SCENARIOS. THAT IS WHERE OUR WORK COMES IN. WE ARE LOOKING AT THE DETERMINISTIC QUALITIES, TO HOPEFULLY ENSURE THE SCALABILITY OF THESE PROGRAMS BY REDUCING THE NUMBER OF SCENARIOS BENEFICIAL TO GIVE THE TREATMENT. GIVEN THAT WE HAVE MARKETS IN THE U.S., WE MAKE A LOT OF DECISIONS, AND IDEALLY, WE WOULD BE ABLE TO SOLVE THE OPTIMAL POWER FLOW, ALL DECISIONS, WERE ALL UNCERTAIN SCENARIOS, WITH THE QUALITY CONTROL PROBLEM. I WOULD BE VERY DIFFICULT TO SOLVE. QUITE REASONABLY, WE BREAK IT DOWN INTO MULTIPLE STEPS. WE MAKE APPROXIMATIONS. WE MAKE APPROXIMATIONS OFF-LINE, AND THEN WE CHECK THESE APPROXIMATIONS AFTER-THE-FACT AND WE MAKE WHAT I'M CALLING UNECONOMICAL ADJUSTMENTS. I'M CALLING THEM THAT BECAUSE USUALLY BY THIS STAGE, WE HAVE FIXED A LOT OF DECISIONS AND ARE ONLY FOCUSED ON LOCALLY OPTIMAL DECISIONS, WHICH IS ECONOMICAL, WHICH YOU COULD HAVE FOUND IF YOU HAD CO-OPTED EVERYTHING IN ONE STAGE. THE APPROACH I WILL PRESENT TO YOU TODAY FITS INTO EACH OF THE THREE CATEGORIES. FIRST, I WILL BE TALKING ABOUT AN OFF-LINE RESERVE ZONE DETERMINATION PROCESS, WHICH IS, IN ORDER TO BETTER ANTICIPATE WHAT THE TRANSITION BOTTLENECKS WILL BE IN THE NEAR FUTURE, I WILL TALK ABOUT THE BASE RESERVE REQUIREMENT WE ARE VETTING WITH MORE DETAILS TO FOLLOW. THEN, ONE OF THE TOOLS OPERATORS HAVE AFTER THEY HAVE A SOLUTION AND HAVE PERFORMED A CONTINGENCY ANALYSIS, IF THEY IDENTIFY GENERATORS BEHIND A TRANSITION BOTTLENECKS. THEY DISQUALIFY THE GENERATOR FROM THE RESERVE IN TODAY'SOLUTION COME BASICALLY ENSURING THE RESERVE REQUIREMENT IS MET BY RESERVE AND OTHER LOCATIONS THAT IS MORE LIKELY DELIVERABLE. WE ARE AND GENTRIFYING THE RESERVE QUALIFICATION. BEGINNING WITH RESERVE SONGS, DYNAMIC IN THE SENSE WE ARE DOING THE MORE FREQUENTLY, BECAUSE ACTUALLY, THE METHODOLOGY FOR DETERMINING RESERVE ZONES IS PRETTY WELL ESTABLISHED AND IS PRETTY STRAIGHTFORWARD, AS WELL. IT INVOLVES TWO STEPS. FIRST IS ANTICIPATING THE TRANSITION BOTTLENECKS WILL BE. THE SECOND IS CLUSTERING TOGETHER -- NODES TOGETHER IN THE SAME ZONE, A SIMILAR EFFECT ON THESE PHYSICAL TRANSITIONS. THE MAIN IDEA BEING IF I WERE TO LOSE A GENERATOR IN ONE'S OWN, I COULD REPLACE THE INJECTION WITH A SMALL CHANGE IN FLOW ON THESE CRITICAL MINDS. A PROCESS, THE BASIC IDEA WELL DOCUMENTED, THE PRIMARY BENEFIT WE ARE LOOKING FOR IS UPDATING THEM MORE FREQUENTLY. THE MOST I AM AWARE OF THIS, EVERY THREE MONTHS, EXCEPT DURING EMERGENCY SITUATIONS. HOWEVER, IT OCCURRED TO US IT IS VERY UNLIKELY THE TRANSITION BOTTLENECKS WILL BE THE SAME EVERY HOUR OF EVERY DAY FOR THE ENTIRETY OF THAT TIME. I WILL PRESENT A FRAMEWORK WE ARE PROPOSING IN ORDER TO DETERMINE UPDATING THE ZONES MORE FREQUENTLY, STARTING WITH THE TRADITIONAL APPROACH. ON A SEASONAL BASIS, WE AGAIN IDENTIFY THE TRANSITION BOTTLENECKS, USUALLY HEAVILY RELYING ON HISTORICAL INFORMATION, WHERE THE CRITICAL TRANSMISSION WILL BE, TO PARTITION THE ZONES, AND PUTTING THAT IN YOU -- INTO YOUR SCHEDULE. WHAT WE HAVE PROPOSED IS DOING THE SAME THING, HOWEVER UPDATING THE ZONES ON A MORE FREQUENT BASIS. THE WAY WE ARE DOING THAT IS AS SUCH. WE ARE SIMULATING THE OPERATION STATE UNDER A LARGE NUMBER OF SCENARIOS. WE ARE USING SIMULATION AND THE OPERATOR RESPONSE TO EACH SCENARIO. IN DOING SO, WE CAN GAIN PROBABILITY OF THE FLOW ON EACH ONE. YOU CAN SAY THE CRITICAL ONES ARE THOSE HEAVILY UTILIZED, A CERTAIN PERCENTAGE OF THE TIME. THE BENEFIT OF DOING THIS IS THAT SINCE WE ARE DOING IT MORE FREQUENTLY, WE CAN TAKE INTO ACCOUNT THE LATEST WIND FORECASTS, ANTICIPATED SUBMITTED, ETC.. THAT IS THE BENEFIT WE ARE LOOKING FOR. AND WE TEST THESE TWO APPROACHES. WE HAVE A MODIFIED VERSION, WITH 20% OF PENETRATION, AND WE MODELED 12 DIFFERENT DAYS. WE USE A TRADITIONAL ZONE APPROACH TO MIMIC WHAT IS DONE IN THE INDUSTRY, AND WE USE THE SAME ZONES FOR ALL FOUR THOSE DAYS AND WE UP DATED THE ZONES. COMPARING THE SEASONAL ZONE FOR ALL OF THEM TO A DAILY ZONE, UNCONDITIONED TO EACH. WE ALSO COMPARED THESE TWO PROGRAMS USING ONLY 10 SCENARIOS. AFTER WE HAD A SOLUTION FOR ALL OF THESE, WE PERFORMED A CONTINGENCY ANALYSIS USING SIMULATION ON 1007 -- SCENARIOS. FOR EACH OF THOSE 1001 SCENARIOS, WE DID 10 MINUS ONE ANALYSIS ON TOP OF IT. WE WILL STOP FOR A MOMENT AND SEE WHERE WE ARE. WE DO AN OFF-LINE DETERMINATION ZONE. THEN, WE ARE DOING A CONTINGENCY ANALYSIS. THE RESULTS I WILL PRESENT ARE STOPPING RIGHT THERE. THEY ARE HOW THESE SOLUTIONS COMPARE AT THIS POINT WHILE ACKNOWLEDGING OPERATORS STILL HAVE THE ABILITY TO ADJUST THESE AT SOME COST, IF NOT RELIABLE. COMPARE THESE RESULTS TO THE TRADITIONAL SEASONAL APPROACH TO THE DAILY APPROACH, WHERE THIS FIRST AXIS IS OFF FOR THE FORECAST. WE HAVE GOT THIS CONTINGENCY ANALYSIS. DUE TO THE RESERVE QUANTITIES WE SELECTED, THE VIOLATIONS ONLY OCCUR WHEN RESERVE IS NOT DELIVERABLE. THE VIOLATION IS ZERO. PENS ON IF THESE ARE NEGLIGIBLE. NOT ZERO IF THEY ARE NOT DELIVERABLE. UPDATING ZONES ON A DAILY BASIS APPEARS TO BE A GOOD WAY OF IMPROVING THE DELIVERABILITY OF RESERVES. COMPARING THIS, ONE THING TO IMMEDIATE NOTE -- IMMEDIATELY NOTE IS THAT THE RESERVE IS NOT DELIVERABLE. THERE IS A SIMPLE EXPLANATION. IT IS COUSIN -- BECAUSE WE HAVE REMODELED 10 SCENARIOS. WHEN WE EVALUATED AGAINST 1001 SCENARIOS, WE ARE SCHEDULING INTO A SMALL SUBSET OF SCENARIOS BUT IT STILL GOES IN ORDER OF MAGNITUDE, AND THIS IS ONE OF THE CHALLENGES WE ARE FACING, THAT IF WE WANT TO RESUME -- RESOLVE -- SOLVE THESE IN A TIMEFRAME, WE NEED TO USE SCENARIOS WHICH MAY NOT YET US TO WHERE WE WANT DETERMINING IF THE SYSTEM IS RELIABLE AND SECURE. THE LAST THING I WILL SHOW IS COMPARING -- COMBINING THIS APPROACH TO OUR DAILY UPDATES OF ZONES USING THE POLICIES TOGETHER, HAS THE BEST DELIVERABILITY OF ALL OF THE ALTERNATIVES. ONE VERY IMPORTANT THING I HAVE NOT TALKED ABOUT IS THE COST. YOU WILL NOTICED THE PROGRAM HAS THE BEST OUT OF ALL OF THESE. ONE THING WE HAVE TO KEEP IN MIND AND REALIZE IS THAT AN OPERATOR LOOKING AT THE SOLUTION WILL BE OBLIGED, IN ORDER TO ENSURE -- TO WRECK THE SOLUTION SO LIABILITY AND PRESUMABLY THE LESS DELIVERABLE A RESERVE IS, THE MORE EXPENSIVE THE CORRECTION MAY BE. WHICH ONE OF THESE METHODS COMES OUT ON TOP IN THE END -- DEPENDS VERY MUCH ON THAT PROCESS. TO DRIVE THAT POINT HOME, I PUT THE SAME OF -- THE SAME RESULTS WE JUST SAW FOR THOSE METHODS, VIOLATION ON THE Y AXIS, OPERATORS WILL MAKE SOME ADJUSTMENT, AND THAT WILL BRING THE VIOLATION DOWN TO ZERO, BUT THAT WILL HAVE SOME COSTS. IT IS OUR BELIEF HAVING A GOOD APPROACH FOR MAKING THE ADJUSTMENTS IS CRITICAL AND COULD BE IN TERMS OF EVALUATING SEPARATE AND DIFFERENT TYPES OF POLICIES. THIS IS ONGOING WORK WE ARE WORKING ON. IN THE INTEREST OF TIME, FOR THE REMAINING APPROACHES, I WOULD JUST GIVE YOU A HIGH-LEVEL OVERVIEW OF WHAT WE ARE DOING. I WILL NOT GO IN DEPTH INTO THE ONGOING WORK. THE FIRST THING, CONGESTION- BASED RESERVES. THE MOTIVATION FOR THIS IS SOMETHING YOU CAN PRACTICE TODAY WHERE YOU LOOK AT DYNAMICS RESERVE. SAY YOU HAVE TWO ZONES. ZONE A AND ZONE BE. ZONE B IS USUALLY IMPORTING CHROMOSOME A. YOU MIGHT SAY IT IS IN YOUR MARKET SOLUTION, THIS INTERFACE HAPPENS TO BE LIGHTLY EMOTED THAT ZONE B SHOULD BE ABLE TO IMPORT MORE RESERVES THAN WHEN IT IS LOADED. -- HEAVILY LOADED. THIS IS A SIMPLE WAY OF PUTTING IT, BUT MAKING THESE TYPES OF CONSIDERATIONS DYNAMICALLY WITHIN THEIR SCHEDULING MODEL, WE ARE TAKING THE SAME TYPES OF APPROACH AND TRYING TO APPLY IT WITHIN THE ZONE ITSELF. WE ARE DOING THIS THROUGH THE USE -- YOUTH -- USE OF A TRANSMISSION. THIS IS NOT DIFFERENT FROM THE LITERATURE, WHERE AS IT GOES UP, THEY WANT TO POST CAPS ON AT. THE DIFFERENCE IS WHAT WE ARE DOING IS AVID GOES -- AS IT GOES UP, WE ANTICIPATE THE OTHER CAMERAS ARE MORE RESERVES. THERE IS OR WILL ALLOW THE OTHERS ARE TO LEAD -- LEAVE. THE WAY IT WORKS, WE LOOK AS THEY ARE NOW AND WE ANTICIPATE HOW THEY WILL CHANGE AFTER A CONTINGENCY AND RE-DISPATCH THAT CONTINGENCY AFTER THE THINGS OCCUR. WE ARE NOT SERVING THAT BUT WE ARE DOING THAT IN LIMITED SETS TO TRY TO GET A ROUGH IDEA OF HOW STRESSED THE SYSTEM WILL BE FOLLOWING A CONTINGENCY AND WE HAVE A PAPER SUBMITTED, POWER SYSTEMS UNDER REVIEW RIGHT NOW. AN INTERESTING PART IS IN THE RESULTS WHEN THE METHOD WORKS, IT WAS NOT BECAUSE IT WAS IDENTIFYING A GOOD RESERVE. WHEN IT WORKED EFFICIENTLY, IT WAS BECAUSE IT WAS REDUCING THE TRANSMISSION STRESS ON CRITICAL LINES. BY EFFICIENTLY, I MEAN, IN SOME INSTANCES, IT WAS CHEAPER, TO REDUCE ON A CRITICAL LINE, THEN IT WAS TO ACQUIRE ADDITIONAL RESERVES. THESE ARE THE TYPES OF TRADE- OFFS THAT CAN OCCUR IN THE MODEL , ASSUMING YOU SPECIFIED IT IN A REASONABLE WAY. THE LAST THING I WILL TALK ABOUT IS RESERVE QUALIFICATION AND RESERVE DOWNPLAY, SOMETHING IT -- THAT IS DONE DURING THE DAY. AFTER THEY HAVE A SOLUTION. IF YOU IDENTIFY A GENERATOR IS BEHIND THE BOTTLENECK, YOU HAVE THE OPTION TO DISQUALIFY THE RESERVE, COMPOUNDING THE REQUIREMENT SO THAT THE REQUIREMENT HAS TO BE MET BY RESERVE AND OTHER LOCATIONS. THIS IS DONE LARGELY MANUALLY TODAY. WHAT WE ARE WORKING ON IS WHAT WE ARE CALLING GENERALIZED. ONE OF THE BENEFITS WE ARE TARGETING IS THAT WE CAN DO THE MATHEMATICAL PROGRAMMING SO WE CAN PROVIDE THE SUPPORT TOOLS TO HELP OPERATORS MAKE THE BETTER DECISIONS IN THE AREA. ANOTHER OPPORTUNITY WE ARE LOOKING TO LEVERAGE IS APPLYING THESE ON A PERSON AREA BASIS. TO GIVE A QUICK EXAMPLE OF THAT, SAY YOU HAVE SOME RESERVE, IN THIS LEFT AREA HERE, AND THERE IS A BASE FLOW, THE CONTINGENCY, FOR EXAMPLE, BETWEEN THE LEFT AND RIGHT AREA. IF GENERATOR ONE WERE TOO FAIR, AND SUDDENLY WE LOSE A LOT OF FLOW ON THIS TOPLINE, IT MAY BE THE RESERVE CAN NOT ACTIVATED WITHOUT VIOLATING THE TRANSMISSION. IF GENERATOR TWO WERE TO FAIR COP -- FAIL, GENERATOR ONE IS STILL PROVIDING THE COUNTER FLOW ON THE TOP LINE AND IT MAY BE THE CASE IT IS DELIVERED. WHAT WE ARE TRYING TO CAPTURE IT ARE THE INSTANCES WHERE THE RESERVE IS DELIVERABLE 99 TIMES OUT OF 100, BUT NOT JUST FOR THE LAST ONE. WE ARE NOT DISQUALIFYING THEM FOR EVERYTHING, BUT ONLY THE SCENARIOS IN WHICH THE RESERVE IS NOT DELIVERABLE. THIS IS SOMETHING WE ARE PROPOSING TO BE APPLIED. IT IS ONGOING WORK AND CAN BE APPLIED AFTER THE FACT TO ADJUST THE SOLUTION IN ACCORDANCE TO THE TYPES OF POLICIES WE HAVE TODAY, AND WE ARE ALSO CONSIDERING THIS IS SOMETHING THAT CAN BE INDEBTED AS PART OF A DECOMPOSITION -- COMPOSITION ALGORITHM. THAT IS ONGOING WORK. I TALKED ABOUT THREE DIFFERENT WAYS BUT IT IS REALLY MOTIVATED BY THIS. IT IS HARD TO PREDICT TRANSITION BOTTLENECKS. IT MAY BECOME MORE DIFFICULT, AND WHEN WE HAVE HIGHER PENETRATIONS, FURTHERMORE, PROGRAMMING CANNOT CAPTURE ALL THE UNCERTAINTY WITHIN THE TIME LIMITATIONS WE HAVE TO SOLVE THE PROBLEM. DETERMINISTIC REQUIREMENTS CAN BUT THEY MAY NOT BE EFFICIENT. WHAT WE ARE WORKING ON IS BETTER DETERMINISTIC POLICIES, A PORTFOLIO OF THEM THAT CAN BE APPLIED AT DIFFERENT POINTS AHEAD OF THE SCHEDULING PROCESS IN ORDER TO NOT ONLY IMPROVE THE DETERMINE A STICK APPROACHES, BUT ALSO FOR THE AVAILABILITY OF PROGRAMS. ONE OF THE THINGS WE ARE INTERESTED IN IS HOW WE VIEW -- VIEW THESE POLICIES TO WORK TO SET UP SCENARIOS BEING OMITTED. HOW DO WE DESIGN THE POLICY TO SPECIFICALLY LOOK AT THOSE. ALSO, POTENTIAL APPLICATION, HORIZONTAL ALGORITHMS. OFF ABOUT THAT YESTERDAY, WE ARE CURRENTLY IN THE PROCESS, AND, I WAS VERY HAPPY TO PRESENT TO YOU TODAY. THANK YOU FOR YOUR ATTENTION. ONE OF THE THINGS WE ARE LOOKING FOR AND HOPING FOR IS THAT WE CAN GET FEEDBACK ON THIS. IF YOU HAVE COMMENTS QUESTIONS OR CONCERNS OR WANT ANY MORE INFORMATION, I WOULD INVITE YOU AND ASK YOU TO TALK TO EACH ONE OF US OFF-LINE. AT THIS POINT, I CAN GO TO QUESTIONS -- ONLINE. [APPLAUSE]

>> GREAT PRESENTATION. CAN YOU DESCRIBE A LITTLE BIT MORE -- YOU DISCUSSED THE ISSUE OF LOCATIONAL DELIVERABILITY. YOU ALSO DISCUSSED BRIEFLY THE ISSUE OF WHERE THE CONTINGENCY OCCURRED. CAN YOU DISCUSS ABOUT WHAT YOU ARE USING AND WHAT YOU RECOMMEND FOR THE SELECTION OF THOSE ISSUES AND HOW BROAD YOU THINK THAT NEEDS TO GO FOR THE OTHER SIDE OF THE PATRIOT L WHERE IT COMES FROM, CHARACTERIZING WHAT WE DO. THE LARGEST GENERATOR. MAYBE THAT SAYS [INDISCERNIBLE] BUT IT GETS MORE COLLOCATED. QWEST COULD YOU ELABORATE ON THE DEMAND SIDE?

>> WHAT CONTINGENCY DO YOU MODIFY [INDISCERNIBLE]

>> OK. I WILL ANSWER THAT QUESTION IN A COUPLE OF WAYS. FIRST OFF, MODEL CONSTRAINTS FOR THE RECOURSE CAN BE APPLIED TO EACH SCENARIO. IF YOU ARE LOOKING AT DETERMINISTIC POLICIES, THEY ACTUALLY DO A RESERVE SHARING ANALYSIS. 48 HOURS AT OF TIME, THE DAY BEFORE, THEY DO A SIMULATION OF EACH CONTINGENCY, AND THEY TRIED TO EVALUATE HOW MUCH RESERVE THEY CAN IMPORT IN EACH SCENARIO. THEY ARE DOING SO, AND PERHAPS THE DOWNSIDE OF THAT IS THAT THEY ARE DOING IT DAYS AHEAD A TIME. IT IS HARD TO KNOW. HOWEVER, I WOULD CLASSIFY, THIS IS AS A LIMITATION OF THE POLICIES WE HAVE TODAY. ONE WAY OF POTENTIALLY ADDRESSING THAT IS TO NOT HAVE RESERVE ZONES OF THIS FORM BUT OVERLAPPING RESERVES. A SEPARATE RESERVE SEWN -- ZONE FOR CONTINGENCIES. OR OVERLAPPING ZONE. WORRY REQUIREMENT FOR THE CONTINGENCY WHERE YOU HAVE A RESPONSE FOR EACH OF THOSE. THAT WOULD BE ONE WAY OF ADDRESSING IT.

>> WE HAVE TIME FOR ONE OR TWO MORE.

>> WHEN YOU ARE ABOUT YOUR VIOLATIONS, --

>> [INAUDIBLE]

>> WELL, YOU COULD DO WHATEVER YOU WANT. YOU COULD DO IT IN GENERAL. FOR OUR PARTICULAR CASE, YOU ONLY HAVE ONE CAPACITY LIMIT ON EACH ONE.

>> [INDISCERNIBLE]

>> IT DEPENDS ON HOW YOU WANT TO DO IT. THE TRADITIONAL WAY IS TO HAVE -- [INAUDIBLE]

>> THE POINT I'M SAYING IS WHAT WE HAVE SEEN IN MY EXPERIENCE, THE CONTINGENCY RESERVE DEPLOYMENT IS AN AVERAGE 14 MINUTES -- REGULATION DEPLOYMENTS [INDISCERNIBLE] I LIKE TO GO ON THE OTHER SIDE OF THE FENCE AND SAY, [INAUDIBLE] ENGINEERING ALL THE WAY UP TO THE PLANNING, IS THIS SOMETHING WE HAVE TO GO THROUGH 415 OR 20 MINUTES, JUST LET IT GO. THESE ARE SO FAST AND SO ACCURATE AND PICKING UP IN THE NEXT FIVE MINUTES -- IS THERE ANOTHER OPTION WE COULD TAKE, AND SAY, LEAD -- LET'S HAVE AN IMMEDIATE [INDISCERNIBLE] I DO APPRECIATE THE WORK WE -- YOU HAVE DONE. SOMETIMES YOU HAVE TO STEP BACK.

>> I'M EAGER TO ANSWER THE QUESTION. MY RESPONSE WITH THE MICROPHONE OVER THERE IS THAT CERTAINLY, WHAT WE ARE PROPOSING DEPENDS VERY MUCH ON THE SITUATION AND WHAT ELSE YOU ARE DOING. IF YOU ARE OPERATING RIGHT AT THE LIMIT AND IF YOU WERE TO TAKE ONE MORE STEP, YOU WOULD BE IN VIOLATION OF WHAT YOU CHOSE TO BE YOUR SECURITY CRITERION ANYONE TO OPERATE CLOSE TO THE LIMIT ON SOMETHING YOU ARE ABOUT DOING TO OPERATE MORE [INDISCERNIBLE] AND THIS IS AN ALTERNATIVE, PERHAPS YOU COULD OPERATE IN THAT DIRECTION WITHOUT --

>> THE GENERAL FRAMEWORK ON HOW WE FEAR IS UNCERTAINTY AND CONTINGENCIES, NOT AS THE SHORT- TERM PROCESS BUT THE OFF-LINE PROCESS HE IS MEANT TO -- MENTIONING. YOU CAN START SAYING, I WILL HAVE AMPLE CAPACITY AVAILABLE SUCH THAT IF A CONTINGENCY HAPPENS IN REAL TIME, I AM FINE. YOU CAN DO THAT BUT THEN YOU ARE INEFFICIENT. THE WHOLE PROCESS STARTS FROM THE VERY BEGINNING IN TERMS OF IMPROVING OUR ABILITY TO MODEL UNCERTAINTIES SO WE ARE MORE EFFICIENT FROM THE GET-GO TO THE REAL-TIME PROCESS BY IMPROVING OUR OPERATIONS. THIS IS BOTH A COMMENT AS WELL AS A SUBSTITUTE TO THE OTHER MESSAGE THAT YOU CAN IDENTIFY ALL OF THESE DIFFERENT PROCESSES WE USE TO ENSURE OUR SYSTEM IS FEASIBLE GETS US THERE, BUT IT IS NOT EFFICIENT.

>> I WILL PARAPHRASE. HE SAID IF YOU WANT TO ENSURE RELIABILITY, I CAN DO THAT AND IT IS EASY. THE DIFFICULTY IS TRYING TO DO IT ECONOMICALLY. THAT IS WHAT A LOT OF US ARE TRYING TO PUSH THE BOUNDARIES ON. QWEST THANK YOU. --

>> THANK YOU. [APPLAUSE] .