WEBVTT 00:00:00.000 --> 00:00:01.992 align:middle line:90% [SQUEAKING] 00:00:01.992 --> 00:00:03.486 align:middle line:90% [RUSTLING] 00:00:03.486 --> 00:00:17.070 align:middle line:90% [CLICKING] 00:00:17.070 --> 00:00:18.570 align:middle line:84% RICHARD DE NEUFVILLE: At this point, 00:00:18.570 --> 00:00:21.100 align:middle line:84% I'm going to talk about the third element of the garage 00:00:21.100 --> 00:00:21.600 align:middle line:90% case. 00:00:21.600 --> 00:00:23.580 align:middle line:84% We already started with the idea of what 00:00:23.580 --> 00:00:26.040 align:middle line:84% the deterministic engineering design was, 00:00:26.040 --> 00:00:29.130 align:middle line:84% then we saw how that analysis led 00:00:29.130 --> 00:00:32.610 align:middle line:84% to a wrong estimate of the value and, in fact, the wrong design 00:00:32.610 --> 00:00:36.220 align:middle line:84% once we accounted for the reality of uncertainty. 00:00:36.220 --> 00:00:39.240 align:middle line:84% Now let's think how we can be proactive about this design, 00:00:39.240 --> 00:00:41.940 align:middle line:84% and we're going to do that by introducing flexibility. 00:00:41.940 --> 00:00:44.660 align:middle line:90% What does that mean? 00:00:44.660 --> 00:00:47.900 align:middle line:84% The flexibility is the ability to adjust the project 00:00:47.900 --> 00:00:52.100 align:middle line:84% to the actual needs, according to how the future develops. 00:00:52.100 --> 00:00:56.150 align:middle line:84% In particular, it means that if the future 00:00:56.150 --> 00:00:59.013 align:middle line:84% is very positive in terms of demand 00:00:59.013 --> 00:01:00.680 align:middle line:84% and you'd like to have greater capacity, 00:01:00.680 --> 00:01:06.460 align:middle line:84% that you have the capability of adding that if needed. 00:01:06.460 --> 00:01:09.280 align:middle line:84% It also presumes that we are managing this project. 00:01:09.280 --> 00:01:11.620 align:middle line:84% We just don't create it and say, there. 00:01:11.620 --> 00:01:12.910 align:middle line:90% It's off on its own. 00:01:12.910 --> 00:01:14.260 align:middle line:90% It's on autopilot. 00:01:14.260 --> 00:01:14.980 align:middle line:90% No. 00:01:14.980 --> 00:01:17.440 align:middle line:84% We have a system where we manage it appropriately. 00:01:17.440 --> 00:01:19.240 align:middle line:84% So it's very important to understand 00:01:19.240 --> 00:01:22.420 align:middle line:84% that if we want to deliver the value of a design, 00:01:22.420 --> 00:01:24.250 align:middle line:84% it's not only what we do at the beginning, 00:01:24.250 --> 00:01:27.490 align:middle line:84% but how we manage the project as it encounters 00:01:27.490 --> 00:01:29.650 align:middle line:84% different situations and has to deal 00:01:29.650 --> 00:01:32.170 align:middle line:84% with different possibilities and opportunities. 00:01:32.170 --> 00:01:34.480 align:middle line:84% How do we do this for the garage? 00:01:34.480 --> 00:01:37.030 align:middle line:84% For the garage case, what we do is we 00:01:37.030 --> 00:01:39.430 align:middle line:84% enable the addition of more capacity. 00:01:39.430 --> 00:01:43.090 align:middle line:84% That is, it has to have the strength to carry more loads, 00:01:43.090 --> 00:01:43.840 align:middle line:90% it has-- 00:01:43.840 --> 00:01:46.840 align:middle line:84% if you want fat columns with more strength on them, 00:01:46.840 --> 00:01:48.970 align:middle line:84% it has to have bigger foundations 00:01:48.970 --> 00:01:51.790 align:middle line:84% to carry the loads on the soil, and it 00:01:51.790 --> 00:01:55.540 align:middle line:84% has to have a building system that you can easily 00:01:55.540 --> 00:02:00.010 align:middle line:84% add to, in this case a precast situation. 00:02:00.010 --> 00:02:01.960 align:middle line:84% You build it offsite, can bring it in, 00:02:01.960 --> 00:02:06.130 align:middle line:84% and can expand it with a minimal closure of the garage 00:02:06.130 --> 00:02:07.940 align:middle line:90% during the expansion period. 00:02:07.940 --> 00:02:12.280 align:middle line:84% So then you have the ability to add floors and what have you. 00:02:12.280 --> 00:02:14.590 align:middle line:84% Now, let's think about how we're going 00:02:14.590 --> 00:02:18.400 align:middle line:84% and analyze this with the simulation process. 00:02:18.400 --> 00:02:20.360 align:middle line:90% Here is a spreadsheet of it. 00:02:20.360 --> 00:02:24.070 align:middle line:84% And this is the one for four floors. 00:02:24.070 --> 00:02:26.080 align:middle line:90% Now, what we have here-- 00:02:26.080 --> 00:02:29.290 align:middle line:84% I won't explain in the detail how it works, 00:02:29.290 --> 00:02:33.430 align:middle line:84% but I'll give you the idea of it, is that we have a forecast 00:02:33.430 --> 00:02:35.860 align:middle line:84% here, and this is a simulated forecast. 00:02:35.860 --> 00:02:38.560 align:middle line:84% That is, sometimes the top number of the forecast 00:02:38.560 --> 00:02:40.870 align:middle line:84% will be up, sometimes it'll be down. 00:02:40.870 --> 00:02:42.790 align:middle line:90% It'll be up and down over time. 00:02:42.790 --> 00:02:45.640 align:middle line:90% And that's what we observe. 00:02:45.640 --> 00:02:49.270 align:middle line:84% Then we think about how we would exercise, 00:02:49.270 --> 00:02:52.240 align:middle line:84% the flexibility that's available in terms 00:02:52.240 --> 00:02:54.010 align:middle line:90% of intelligent management. 00:02:54.010 --> 00:02:56.980 align:middle line:84% So if the demand is not increasing, 00:02:56.980 --> 00:02:58.660 align:middle line:90% we do not add to capability. 00:02:58.660 --> 00:03:01.310 align:middle line:84% If the-- extra floors, in particular. 00:03:01.310 --> 00:03:03.580 align:middle line:84% If the demand however does increase, 00:03:03.580 --> 00:03:05.980 align:middle line:84% then it looks like it'd be valuable to have 00:03:05.980 --> 00:03:09.250 align:middle line:90% more floors on it, we add them. 00:03:09.250 --> 00:03:11.950 align:middle line:84% We've embedded in here through what's 00:03:11.950 --> 00:03:14.890 align:middle line:84% called if statements that looks at what's been happening 00:03:14.890 --> 00:03:17.140 align:middle line:84% and what would be the intelligent thing to do. 00:03:17.140 --> 00:03:19.450 align:middle line:84% And for each simulation that we run, 00:03:19.450 --> 00:03:24.050 align:middle line:84% we factor that in and calculate the value. 00:03:24.050 --> 00:03:27.130 align:middle line:84% So we do this for thousands of times, 00:03:27.130 --> 00:03:29.500 align:middle line:84% a couple thousand in this particular analysis. 00:03:29.500 --> 00:03:33.520 align:middle line:84% And we look at-- sometimes you expand because it's desirable, 00:03:33.520 --> 00:03:34.960 align:middle line:90% sometimes you do not expand. 00:03:34.960 --> 00:03:37.240 align:middle line:84% We look at the combination and figure out 00:03:37.240 --> 00:03:39.920 align:middle line:90% the overall expected value. 00:03:39.920 --> 00:03:43.000 align:middle line:90% So where does that lead us? 00:03:43.000 --> 00:03:47.710 align:middle line:84% It leads us to-- as we look at all the possibilities, one 00:03:47.710 --> 00:03:49.600 align:middle line:84% story, two story, and so forth with 00:03:49.600 --> 00:03:52.570 align:middle line:84% and without the flexibility, that the optimal design 00:03:52.570 --> 00:03:54.670 align:middle line:84% in this particular case with these numbers 00:03:54.670 --> 00:03:58.990 align:middle line:84% is not six stories, which was a deterministic design, not 00:03:58.990 --> 00:04:03.760 align:middle line:84% the five stories, which is the risk averse, avoiding 00:04:03.760 --> 00:04:07.240 align:middle line:84% the downside approach where you recognize the uncertainty, 00:04:07.240 --> 00:04:08.560 align:middle line:90% but it's four stories. 00:04:08.560 --> 00:04:13.730 align:middle line:84% But four stories with the ability to expand when needed. 00:04:13.730 --> 00:04:15.280 align:middle line:90% So observe what's happening. 00:04:15.280 --> 00:04:17.260 align:middle line:84% You've built it smaller so it didn't 00:04:17.260 --> 00:04:19.899 align:middle line:84% cost as much as having five or six stories 00:04:19.899 --> 00:04:22.600 align:middle line:84% even though you had strong foundations. 00:04:22.600 --> 00:04:25.420 align:middle line:84% So it's less exposure to the downside, 00:04:25.420 --> 00:04:28.780 align:middle line:84% but you've built it so that you get five, six, seven, or eight 00:04:28.780 --> 00:04:31.850 align:middle line:90% stories if the demand occurred. 00:04:31.850 --> 00:04:36.010 align:middle line:84% So here is the summary of the design results. 00:04:36.010 --> 00:04:41.200 align:middle line:84% So that on the one hand, we have the different possibilities 00:04:41.200 --> 00:04:44.380 align:middle line:84% and their performance over time, and we 00:04:44.380 --> 00:04:47.060 align:middle line:90% have the value at the end. 00:04:47.060 --> 00:04:52.390 align:middle line:84% So what we observe here is that the four story building 00:04:52.390 --> 00:04:55.000 align:middle line:90% is the most productive one. 00:04:55.000 --> 00:04:59.020 align:middle line:84% And we might have thought that the deterministic design 00:04:59.020 --> 00:05:01.630 align:middle line:84% with a value of 6 million was better, 00:05:01.630 --> 00:05:05.980 align:middle line:84% but that result is a sham because given the uncertainty, 00:05:05.980 --> 00:05:09.310 align:middle line:84% we will never get that result as the expected value. 00:05:09.310 --> 00:05:11.200 align:middle line:84% Its expected value will be lower. 00:05:11.200 --> 00:05:13.390 align:middle line:90% And, in fact, in the case unc-- 00:05:13.390 --> 00:05:15.760 align:middle line:84% the reality case with the uncertainty 00:05:15.760 --> 00:05:18.490 align:middle line:84% is we'll get only about 3 million 00:05:18.490 --> 00:05:22.190 align:middle line:84% with it as the value as compared to what we get. 00:05:22.190 --> 00:05:25.900 align:middle line:84% So let's think about what's happening here. 00:05:25.900 --> 00:05:29.150 align:middle line:84% We build it smaller, less cost at the beginning, 00:05:29.150 --> 00:05:30.950 align:middle line:90% less possibility of loss. 00:05:30.950 --> 00:05:32.590 align:middle line:84% On the other hand, we can build it up 00:05:32.590 --> 00:05:36.160 align:middle line:84% as much as seems appropriate in the particular conditions 00:05:36.160 --> 00:05:39.160 align:middle line:84% without a-- we have favorable demand 00:05:39.160 --> 00:05:41.960 align:middle line:90% and we can make more value. 00:05:41.960 --> 00:05:45.880 align:middle line:84% So let's look about what it is, how 00:05:45.880 --> 00:05:50.930 align:middle line:84% we think about this overall for this particular design case. 00:05:50.930 --> 00:05:55.400 align:middle line:84% So we need to have a multidimensional evaluation. 00:05:55.400 --> 00:05:57.200 align:middle line:90% Why multidimensional? 00:05:57.200 --> 00:06:00.380 align:middle line:84% Because since we have a distribution, 00:06:00.380 --> 00:06:03.710 align:middle line:84% one number, one value for it is not 00:06:03.710 --> 00:06:05.000 align:middle line:90% describing what's happening. 00:06:05.000 --> 00:06:08.660 align:middle line:84% You care not only about the average, you care about, 00:06:08.660 --> 00:06:11.030 align:middle line:84% how bad can it get, how good can it 00:06:11.030 --> 00:06:12.710 align:middle line:90% get, what's the possibilities? 00:06:12.710 --> 00:06:15.320 align:middle line:90% In this particular case, that-- 00:06:15.320 --> 00:06:19.430 align:middle line:84% we see that the several factors that we're 00:06:19.430 --> 00:06:23.930 align:middle line:84% looking at in terms of the initial investment 00:06:23.930 --> 00:06:26.180 align:middle line:84% that we-- how much money do we have to put down 00:06:26.180 --> 00:06:27.500 align:middle line:90% to start the process? 00:06:27.500 --> 00:06:30.890 align:middle line:84% It is lower for the flexible design 00:06:30.890 --> 00:06:32.540 align:middle line:90% because we build it smaller. 00:06:32.540 --> 00:06:35.090 align:middle line:84% So what's the expected value of it? 00:06:35.090 --> 00:06:37.170 align:middle line:84% It's very much larger than otherwise 00:06:37.170 --> 00:06:39.170 align:middle line:84% because we can take advantage of the good things 00:06:39.170 --> 00:06:40.830 align:middle line:90% and avoid the bad things. 00:06:40.830 --> 00:06:45.530 align:middle line:84% And in that same way, the maximum value we can get 00:06:45.530 --> 00:06:47.570 align:middle line:84% is larger, the minimum value that we 00:06:47.570 --> 00:06:50.060 align:middle line:84% might be forced to accept is smaller 00:06:50.060 --> 00:06:53.130 align:middle line:90% so that it's good in all cases. 00:06:53.130 --> 00:06:56.880 align:middle line:90% So everything is better. 00:06:56.880 --> 00:06:58.960 align:middle line:90% How did that happen? 00:06:58.960 --> 00:07:02.290 align:middle line:84% We have not had a technological breakthrough in this case. 00:07:02.290 --> 00:07:06.240 align:middle line:84% It is not a more sophisticated or advanced design. 00:07:06.240 --> 00:07:12.240 align:middle line:84% It is because we have changed the design approach fro-- 00:07:12.240 --> 00:07:14.850 align:middle line:84% we've changed the framing of the problem that 00:07:14.850 --> 00:07:17.610 align:middle line:84% instead of going to, say, we're designing it 00:07:17.610 --> 00:07:21.990 align:middle line:84% for a fixed view of the future, and instead 00:07:21.990 --> 00:07:24.990 align:middle line:84% of focusing on this mythical and this fairy tale future that, 00:07:24.990 --> 00:07:27.480 align:middle line:84% in fact, will not be the exact case, 00:07:27.480 --> 00:07:33.060 align:middle line:84% we changed from that to managing the realistic uncertainties 00:07:33.060 --> 00:07:36.330 align:middle line:84% of life by having the flexible design 00:07:36.330 --> 00:07:39.510 align:middle line:84% and, therefore, can adjust to the conditions 00:07:39.510 --> 00:07:44.340 align:middle line:84% and, in that way, have an overall better solution. 00:07:44.340 --> 00:07:47.160 align:middle line:84% Now, in this case, the flexible design 00:07:47.160 --> 00:07:50.160 align:middle line:84% was better on four counts that, as I mentioned, 00:07:50.160 --> 00:07:52.240 align:middle line:90% that may not always happen. 00:07:52.240 --> 00:07:55.620 align:middle line:84% Sometimes the situation is not so uniquely favorable, 00:07:55.620 --> 00:07:59.010 align:middle line:84% but it's the kind of thing can happen because it's built in 00:07:59.010 --> 00:08:02.370 align:middle line:84% to the way of approaching it as less downside, 00:08:02.370 --> 00:08:05.220 align:middle line:84% more upside, lower cost than beginning. 00:08:05.220 --> 00:08:08.870 align:middle line:84% Overall, it's a win-win at less cost. 00:08:08.870 --> 00:08:14.000 align:middle line:90%